shark Namespace Reference

AbstractMultiObjectiveOptimizer. More...

Namespaces
	blas
	random
	statistics
	tags
	Tags are empty types which can be used as a function argument.

Classes
class	AbsoluteLoss
	absolute loss More...
class	AbstractBudgetMaintenanceStrategy
	This is the abstract interface for any budget maintenance strategy. More...
class	AbstractClustering
	Base class for clustering. More...
class	AbstractConstraintHandler
	Implements the base class for constraint handling. More...
class	AbstractCost
	Cost function interface. More...
class	AbstractKernelFunction
	Base class of all Kernel functions. More...
class	AbstractLinearSvmTrainer
	Super class of all linear SVM trainers. More...
class	AbstractLineSearchOptimizer
	Basis class for line search methods. More...
class	AbstractLoss
	Loss function interface. More...
class	AbstractMetric
class	AbstractModel
	Base class for all Models. More...
class	AbstractMultiObjectiveOptimizer
	base class for abstract multi-objective optimizers for arbitrary search spaces. More...
class	AbstractNearestNeighbors
	Interface for Nearest Neighbor queries. More...
class	AbstractObjectiveFunction
	Super class of all objective functions for optimization and learning. More...
class	AbstractOptimizer
	An optimizer that optimizes general objective functions. More...
class	AbstractSingleObjectiveOptimizer
	Base class for all single objective optimizer. More...
class	AbstractStoppingCriterion
	Base class for stopping criteria of optimization algorithms. More...
class	AbstractSvmTrainer
	Super class of all kernelized (non-linear) SVM trainers. More...
class	AbstractTrainer
	Superclass of supervised learning algorithms. More...
class	AbstractUnsupervisedTrainer
	Superclass of unsupervised learning algorithms. More...
class	AbstractWeightedTrainer
	Superclass of weighted supervised learning algorithms. More...
class	AbstractWeightedUnsupervisedTrainer
	Superclass of weighted unsupervised learning algorithms. More...
struct	Ackley
	Convex quadratic benchmark function with single dominant axis. More...
class	Adam
	Adaptive Moment Estimation Algorithm (ADAM) More...
struct	AdditiveEpsilonIndicator
	Implements the Additive approximation properties of sets. More...
class	ARDKernelUnconstrained
	Automatic relevance detection kernel for unconstrained parameter optimization. More...
class	BarsAndStripes
	Generates the Bars-And-Stripes problem. In this problem, a 4x4 image has either rows or columns of the same value. More...
class	BaseDCNonDominatedSort
	Divide-and-conquer algorithm for non-dominated sorting. More...
struct	Batch
	class which helps using different batch types More...
struct	Batch< blas::vector< T > >
	specialization for vectors which should be matrices in batch mode! More...
struct	Batch< detail::MatrixRowReference< M > >
struct	Batch< shark::blas::compressed_vector< T > >
	specialization for ublas compressed vectors which are compressed matrices in batch mode! More...
struct	Batch< WeightedDataPair< DataType, WeightType > >
struct	BatchTraits
struct	BatchTraits< blas::compressed_matrix< T > >
struct	BatchTraits< blas::dense_matrix_adaptor< T, blas::row_major > >
struct	BatchTraits< blas::matrix< T > >
struct	BatchTraits< WeightedDataBatch< DataType, WeightType > >
class	BFGS
	Broyden, Fletcher, Goldfarb, Shannon algorithm for unconstraint optimization. More...
class	BiasSolver
class	BiasSolverSimplex
class	BinaryLayer
	Layer of binary units taking values in {0,1}. More...
class	BinaryTree
	Super class of binary space-partitioning trees. More...
class	BipolarLayer
	Layer of bipolar units taking values in {-1,1}. More...
struct	BitflipMutator
	Bitflip mutation operator. More...
class	BlockMatrix2x2
	SVM regression matrix. More...
struct	BoxBasedShrinkingStrategy
	Takes q boxx constrained QP-type problem and implements shrinking on it. More...
class	BoxConstrainedProblem
	Quadratic program with box constraints. More...
class	BoxConstrainedShrinkingProblem
class	BoxConstraintHandler
class	BoxedSVMProblem
	Boxed problem for alpha in [lower,upper]^n and equality constraints. More...
class	CachedMatrix
	Efficient quadratic matrix cache. More...
struct	CanBeCalled
	detects whether Functor(Argument) can be called. More...
struct	CanBeCalled< R(*)(T), Argument >
struct	CanBeCalled< R(T), Argument >
class	CARTree
	Classification and Regression Tree. More...
class	Centroids
	Clusters defined by centroids. More...
class	CG
	Conjugate-gradient method for unconstrained optimization. More...
class	Chessboard
	"chess board" problem for binary classification More...
struct	Cigar
	Convex quadratic benchmark function with single dominant axis. More...
class	CigarDiscus
	Convex quadratic benchmark function. More...
struct	CIGTAB1
	Multi-objective optimization benchmark function CIGTAB 1. More...
struct	CIGTAB2
	Multi-objective optimization benchmark function CIGTAB 2. More...
class	CircleInSquare
class	Classifier
	Conversion of real-valued or vector valued outputs to class labels. More...
class	ClusteringModel
	Abstract model with associated clustering object. More...
class	CMA
	Implements the CMA-ES. More...
struct	CMAChromosome
	Models a CMAChromosomeof the elitist (MO-)CMA-ES that encodes strategy parameters. More...
class	CMACMap
	The CMACMap class represents a linear combination of piecewise constant functions. More...
class	CMAIndividual
class	CMSA
	Implements the CMSA. More...
class	CombinedObjectiveFunction
	Linear combination of objective functions. More...
class	ConcatenatedModel
	ConcatenatedModel concatenates two models such that the output of the first model is input to the second. More...
struct	ConstProxyReference
	sets the type of ProxxyReference More...
struct	ConstrainedSphere
	Constrained Sphere function. More...
class	ContrastiveDivergence
	Implements k-step Contrastive Divergence described by Hinton et al. (2006). More...
class	Conv2DModel
	Convolutional Model for 2D image data. More...
class	CrossEntropy
	Error measure for classification tasks that can be used as the objective function for training. More...
class	CrossEntropyMethod
	Implements the Cross Entropy Method. More...
class	CrossValidationError
	Cross-validation error for selection of hyper-parameters. More...
struct	CrowdingDistance
	Implements the Crowding Distance of a pareto front. More...
class	CSvmDerivative
	This class provides two main member functions for computing the derivative of a C-SVM hypothesis w.r.t. its hyperparameters. The constructor takes a pointer to a KernelClassifier and an SvmTrainer, in the assumption that the former was trained by the latter. It heavily accesses their members to calculate the derivative of the alpha and offset values w.r.t. the SVM hyperparameters, that is, the regularization parameter C and the kernel parameters. This is done in the member function prepareCSvmParameterDerivative called by the constructor. After this initial, heavier computation step, modelCSvmParameterDerivative can be called on an input sample to the SVM model, and the method will yield the derivative of the hypothesis w.r.t. the SVM hyperparameters. More...
class	CSVMProblem
	Problem formulation for binary C-SVM problems. More...
class	CSvmTrainer
	Training of C-SVMs for binary classification. More...
class	CVFolds
class	Data
	Data container. More...
class	DataDistribution
	A DataDistribution defines an unsupervised learning problem. More...
class	DataView
	Constant time Element-Lookup for Datasets. More...
class	DiagonalWithCircle
class	DifferenceKernelMatrix
	SVM ranking matrix. More...
struct	DiffPowers
class	DiscreteKernel
	Kernel on a finite, discrete space. More...
class	DiscreteLoss
	flexible loss for classification More...
struct	Discus
	Convex quadratic benchmark function. More...
class	DistantModes
	Creates a set of pattern (each later representing a mode) which than are randomly perturbed to create the data set. The dataset was introduced in Desjardins et al. (2010) (Parallel Tempering for training restricted Boltzmann machines, AISTATS 2010) More...
class	DoublePole
class	DropoutLayer
struct	DTLZ1
	Implements the benchmark function DTLZ1. More...
struct	DTLZ2
	Implements the benchmark function DTLZ2. More...
struct	DTLZ3
	Implements the benchmark function DTLZ3. More...
struct	DTLZ4
	Implements the benchmark function DTLZ4. More...
struct	DTLZ5
	Implements the benchmark function DTLZ5. More...
struct	DTLZ6
	Implements the benchmark function DTLZ6. More...
struct	DTLZ7
	Implements the benchmark function DTLZ7. More...
class	ElitistCMA
	Implements the elitist CMA-ES. More...
struct	ElitistSelection
	Survival selection to find the next parent set. More...
struct	ELLI1
	Multi-objective optimization benchmark function ELLI1. More...
struct	ELLI2
	Multi-objective optimization benchmark function ELLI2. More...
struct	Ellipsoid
	Convex quadratic benchmark function. More...
struct	EmptyState
	Default State of an Object which does not need a State. More...
struct	Energy
	The Energy function determining the Gibbs distribution of an RBM. More...
class	EnergyStoringTemperedMarkovChain
	Implements parallel tempering but also stores additional statistics on the energy differences. More...
class	EpsilonHingeLoss
	Hinge-loss for large margin regression. More...
class	EpsilonSvmTrainer
	Training of Epsilon-SVMs for regression. More...
struct	EPTournamentSelection
	Survival and mating selection to find the next parent set. More...
class	ErrorFunction
	Objective function for supervised learning. More...
class	EvaluationArchive
	Objective function wrapper storing all function evaluations. More...
class	ExactGradient
class	ExampleModifiedKernelMatrix
class	Exception
	Top-level exception class of the shark library. More...
struct	FastSigmoidNeuron
	Fast sigmoidal function, which does not need to compute an exponential function. More...
class	FisherLDA
	Fisher's Linear Discriminant Analysis for data compression. More...
struct	Fonseca
	Bi-objective real-valued benchmark function proposed by Fonseca and Flemming. More...
class	GaussianKernelMatrix
	Efficient special case if the kernel is Gaussian and the inputs are sparse vectors. More...
class	GaussianLayer
	A layer of Gaussian neurons. More...
class	GaussianRbfKernel
	Gaussian radial basis function kernel. More...
class	GaussianTaskKernel
	Special "Gaussian-like" kernel function on tasks. More...
class	GeneralizationLoss
	The generalization loss calculates the relative increase of the validation error compared to the minimum training error. More...
class	GeneralizationQuotient
	SStopping criterion monitoring the quotient of generalization loss and training progress. More...
class	GeneralQuadraticProblem
	Quadratic Problem with only Box-Constraints Let K the kernel matrix, than the problem has the form. More...
class	GibbsOperator
	Implements Block Gibbs Sampling related transition operators for various temperatures. More...
class	GridSearch
	Optimize by trying out a grid of configurations. More...
struct	GSP
	Real-valued benchmark function with two objectives. More...
class	HardClusteringModel
	Model for "hard" clustering. More...
class	HierarchicalClustering
	Clusters defined by a binary space partitioning tree. More...
struct	Himmelblau
	Multi-modal two-dimensional continuous Himmelblau benchmark function. More...
class	HingeLoss
	Hinge-loss for large margin classification. More...
class	HMGSelectionCriterion
class	HuberLoss
	Huber-loss for for robust regression. More...
struct	HypervolumeApproximator
	Implements an FPRAS for approximating the volume of a set of high-dimensional objects. The algorithm is described in. More...
struct	HypervolumeCalculator
	Frontend for hypervolume calculation algorithms in m dimensions. More...
struct	HypervolumeCalculator2D
	Implementation of the exact hypervolume calculation in 2 dimensions. More...
struct	HypervolumeCalculator3D
	Implementation of the exact hypervolume calculation in 3 dimensions. More...
struct	HypervolumeCalculatorMDHOY
	Implementation of the exact hypervolume calculation in m dimensions. More...
struct	HypervolumeCalculatorMDWFG
	Implementation of the exact hypervolume calculation in m dimensions. More...
struct	HypervolumeContribution
	Frontend for hypervolume contribution algorithms in m dimensions. More...
struct	HypervolumeContribution2D
	Finds the smallest/largest Contributors given 2D points. More...
struct	HypervolumeContribution3D
	Finds the hypervolume contribution for points in 3DD. More...
struct	HypervolumeContributionApproximator
	Approximately determines the point of a set contributing the least hypervolume. More...
struct	HypervolumeContributionMD
	Finds the hypervolume contribution for points in MD. More...
struct	HypervolumeIndicator
	Calculates the hypervolume covered by a front of non-dominated points. More...
struct	HypervolumeSubsetSelection2D
	Implementation of the exact hypervolume subset selection algorithm in 2 dimensions. More...
struct	IHR1
	Multi-objective optimization benchmark function IHR1. More...
struct	IHR2
	Multi-objective optimization benchmark function IHR 2. More...
struct	IHR3
	Multi-objective optimization benchmark function IHR3. More...
struct	IHR4
	Multi-objective optimization benchmark function IHR 4. More...
struct	IHR6
	Multi-objective optimization benchmark function IHR 6. More...
class	ImagePatches
	Given a set of images, draws a set of image patches of a given size. More...
class	INameable
	This class is an interface for all objects which can have a name. More...
class	IndexedIterator
	creates an Indexed Iterator, an Iterator which also carries index information using index() More...
class	IndexingIterator
class	IndicatorBasedMOCMA
	Implements the generational MO-CMA-ES. More...
class	IndicatorBasedRealCodedNSGAII
	Implements the NSGA-II. More...
struct	IndicatorBasedSelection
	Implements the well-known indicator-based selection strategy. More...
class	IndicatorBasedSteadyStateMOCMA
	Implements the \((\mu+1)\)-MO-CMA-ES. More...
class	Individual
	Individual is a simple templated class modelling an individual that acts as a candidate solution in an evolutionary algorithm. More...
class	IParameterizable
	Top level interface for everything that holds parameters. More...
class	IRpropMinus
	This class offers methods for the usage of the improved Resilient-Backpropagation-algorithm without weight-backtracking. More...
class	IRpropPlus
	This class offers methods for the usage of the improved Resilient-Backpropagation-algorithm with weight-backtracking. More...
class	IRpropPlusFull
class	ISerializable
	Abstracts serializing functionality. More...
class	IterativeNNQuery
	Iterative nearest neighbors query. More...
class	JaakkolaHeuristic
	Jaakkola's heuristic and related quantities for Gaussian kernel selection. More...
class	KDTree
	KD-tree, a binary space-partitioning tree. More...
class	KernelBudgetedSGDTrainer
	Budgeted stochastic gradient descent training for kernel-based models. More...
struct	KernelClassifier
	Linear classifier in a kernel feature space. More...
class	KernelExpansion
	Linear model in a kernel feature space. More...
class	KernelMatrix
	Kernel Gram matrix. More...
class	KernelMeanClassifier
	Kernelized mean-classifier. More...
class	KernelSGDTrainer
	Generic stochastic gradient descent training for kernel-based models. More...
class	KernelTargetAlignment
	Kernel Target Alignment - a measure of alignment of a kernel Gram matrix with labels. More...
struct	KeyValuePair
	Represents a Key-Value-Pair similar std::pair which is strictly ordered by it's key. More...
class	KHCTree
	KHC-tree, a binary space-partitioning tree. More...
class	LabeledData
	Data set for supervised learning. More...
class	LabeledDataDistribution
	A LabeledDataDistribution defines a supervised learning problem. More...
class	LabelOrder
	This will normalize the labels of a given dataset to 0..N-1. More...
class	LassoRegression
	LASSO Regression. More...
class	LBFGS
	Limited-Memory Broyden, Fletcher, Goldfarb, Shannon algorithm. More...
class	LCTree
	LC-tree, a binary space-partitioning tree. More...
class	LDA
	Linear Discriminant Analysis (LDA) More...
struct	LibSVMSelectionCriterion
	Computes the maximum gian solution. More...
class	LinearClassifier
	Basic linear classifier. More...
class	LinearCSvmTrainer
class	LinearKernel
	Linear Kernel, parameter free. More...
class	LinearModel
	Linear Prediction with optional activation function. More...
struct	LinearNeuron
	Linear activation Neuron. More...
struct	LinearRankingSelection
	Implements a fitness-proportional selection scheme for mating selection that scales the fitness values linearly before carrying out the actual selection. More...
class	LinearRegression
	Linear Regression. More...
class	LinearSAGTrainer
	Stochastic Average Gradient Method for training of linear models,. More...
class	LineSearch
	Wrapper for the linesearch class of functions in the linear algebra library. More...
class	LMCMA
	Implements a Limited-Memory-CMA. More...
struct	LogisticNeuron
	Neuron which computes the Logistic (logistic) function with range [0,1]. More...
class	LogisticRegression
	Trainer for Logistic regression. More...
class	LooError
	Leave-one-out error objective function. More...
class	LooErrorCSvm
	Leave-one-out error, specifically optimized for C-SVMs. More...
class	LRUCache
	Implements an LRU-Caching Strategy for arbitrary Cache-Lines. More...
struct	LZ1
	Multi-objective optimization benchmark function LZ1. More...
struct	LZ2
	Multi-objective optimization benchmark function LZ2. More...
struct	LZ3
	Multi-objective optimization benchmark function LZ3. More...
struct	LZ4
	Multi-objective optimization benchmark function LZ4. More...
struct	LZ5
	Multi-objective optimization benchmark function LZ5. More...
struct	LZ6
	Multi-objective optimization benchmark function LZ6. More...
struct	LZ7
	Multi-objective optimization benchmark function LZ7. More...
struct	LZ8
	Multi-objective optimization benchmark function LZ8. More...
struct	LZ9
class	MarkovChain
	A single Markov chain. More...
class	MarkovPole
struct	MaximumGainCriterion
	Working set selection by maximization of the dual objective gain. More...
struct	MaximumGradientCriterion
	Working set selection by maximization of the projected gradient. More...
class	MaxIterations
	This stopping criterion stops after a fixed number of iterations. More...
class	McPegasos
	Pegasos solver for linear multi-class support vector machines. More...
class	MeanModel
	Calculates the weighted mean of a set of models. More...
class	MergeBudgetMaintenanceStrategy
	Budget maintenance strategy that merges two vectors. More...
class	MergeBudgetMaintenanceStrategy< RealVector >
	Budget maintenance strategy merging vectors. More...
class	MissingFeaturesKernelExpansion
	Kernel expansion with missing features support. More...
class	MissingFeatureSvmTrainer
	Trainer for binary SVMs natively supporting missing features. More...
class	MklKernel
	Weighted sum of kernel functions. More...
class	MNIST
	Reads in the famous MNIST data in possibly binarized form. The MNIST database itself is not included in Shark, this class just helps loading it. More...
class	ModelKernel
	Kernel function that uses a Model as transformation function for another kernel. More...
class	ModifiedKernelMatrix
	Modified Kernel Gram matrix. More...
class	MOEAD
	Implements the MOEA/D algorithm. More...
class	MonomialKernel
	Monomial kernel. Calculates \( \left\langle x_1, x_2 \right\rangle^m_exponent \). More...
class	MultiChainApproximator
	Approximates the gradient by taking samples from an ensemble of Markov chains running in parallel. More...
class	MultiNomialDistribution
	Implements a multinomial distribution. More...
class	MultiObjectiveBenchmark
	Creates a multi-objective Benchmark from a set of given single objective functions. More...
class	MultiTaskKernel
	Special kernel function for multi-task and transfer learning. More...
struct	MultiTaskSample
	Aggregation of input data and task index. More...
class	MultiVariateNormalDistribution
	Implements a multi-variate normal distribution with zero mean. More...
class	MultiVariateNormalDistributionCholesky
	Multivariate normal distribution with zero mean using a cholesky decomposition. More...
struct	MVPSelectionCriterion
	Computes the most violating pair of the problem. More...
class	NearestNeighborModel
	NearestNeighbor model for classification and regression. More...
class	NearestNeighborModel< InputType, unsigned int >
class	NegativeAUC
	Negative area under the curve. More...
class	NegativeGaussianProcessEvidence
	Evidence for model selection of a regularization network/Gaussian process. More...
class	NegativeLogLikelihood
	Computes the negative log likelihood of a dataset under a model. More...
class	NegativeWilcoxonMannWhitneyStatistic
	Negative Wilcoxon-Mann-Whitney statistic. More...
class	NestedGridSearch
	Nested grid search. More...
class	NeuronLayer
class	NonMarkovPole
	Objective function for single and double non-Markov poles. More...
class	NormalDistributedPoints
	Generates a set of normally distributed points. More...
class	NormalizeComponentsUnitInterval
	Train a model to normalize the components of a dataset to fit into the unit inverval. More...
class	NormalizeComponentsUnitVariance
	Train a linear model to normalize the components of a dataset to unit variance, and optionally to zero mean. More...
class	NormalizeComponentsWhitening
	Train a linear model to whiten the data. More...
class	NormalizeComponentsZCA
	Train a linear model to whiten the data. More...
class	NormalizedKernel
	Normalized version of a kernel function. More...
class	NormalizeKernelUnitVariance
	Determine the scaling factor of a ScaledKernel so that it has unit variance in feature space one on a given dataset. More...
class	Normalizer
	"Diagonal" linear model for data normalization. More...
struct	NormalizerNeuron
struct	NSGA3Indicator
class	OneClassSvmTrainer
	Training of one-class SVMs. More...
class	OneNormRegularizer
	One-norm of the input as an objective function. More...
struct	OnePointCrossover
	Implements one-point crossover. More...
class	OneVersusOneClassifier
	One-versus-one Classifier. More...
class	OptimizationTrainer
	Wrapper for training schemes based on (iterative) optimization. More...
class	PamiToy
struct	PartiallyMappedCrossover
	Implements partially mapped crossover. More...
class	PartlyPrecomputedMatrix
	Partly Precomputed version of a matrix for quadratic programming. More...
class	PCA
	Principal Component Analysis. More...
class	Pegasos
	Pegasos solver for linear (binary) support vector machines. More...
struct	PenalizingEvaluator
	Penalizing evaluator for scalar objective functions. More...
class	Perceptron
	Perceptron online learning algorithm. More...
class	PointSearch
	Optimize by trying out predefined configurations. More...
class	PointSetKernel
	Normalized version of a kernel function. More...
class	PolynomialKernel
	Polynomial kernel. More...
struct	PolynomialMutator
	Polynomial mutation operator. More...
class	PopulationBasedStepSizeAdaptation
	Step size adaptation based on the success of the new population compared to the old. More...
class	PrecomputedMatrix
	Precomputed version of a matrix for quadratic programming. More...
class	ProductKernel
	Product of kernel functions. More...
class	ProjectBudgetMaintenanceStrategy
	Budget maintenance strategy that projects a vector. More...
class	ProjectBudgetMaintenanceStrategy< RealVector >
	Budget maintenance strategy that projects a vector. More...
class	ProxyIterator
	Creates an iterator which reinterpretes an object as a range. More...
class	QpBoxLinear
	Quadratic program solver for box-constrained problems with linear kernel. More...
class	QpConfig
	Super class of all support vector machine trainers. More...
class	QpMcBoxDecomp
class	QpMcLinear
	Generic solver skeleton for linear multi-class SVM problems. More...
class	QpMcLinearADM
	Solver for the multi-class SVM with absolute margin and discriminative maximum loss. More...
class	QpMcLinearATM
	Solver for the multi-class SVM with absolute margin and total maximum loss. More...
class	QpMcLinearATS
	Solver for the multi-class SVM with absolute margin and total sum loss. More...
class	QpMcLinearCS
	Solver for the multi-class SVM by Crammer & Singer. More...
class	QpMcLinearLLW
	Solver for the multi-class SVM by Lee, Lin & Wahba. More...
class	QpMcLinearMMR
	Solver for the multi-class maximum margin regression SVM. More...
class	QpMcLinearReinforced
	Solver for the "reinforced" multi-class SVM. More...
class	QpMcLinearWW
	Solver for the multi-class SVM by Weston & Watkins. More...
class	QpMcSimplexDecomp
struct	QpSolutionProperties
	properties of the solution of a quadratic program More...
class	QpSolver
	Quadratic program solver. More...
class	QpSparseArray
	specialized container class for multi-class SVM problems More...
struct	QpStoppingCondition
	stopping conditions for quadratic programming More...
class	RadiusMarginQuotient
	radius margin quotions for binary SVMs More...
class	RankingSvmTrainer
	Training of an SVM for ranking. More...
class	RBFLayer
	Implements a layer of radial basis functions in a neural network. More...
class	RBM
	stub for the RBM class. at the moment it is just a holder of the parameter set and the Energy. More...
class	RealCodedNSGAIII
	Implements the NSGA-III. More...
struct	RealSpace
	The RealSpace can't be enumerated. Infinite values are just too much. More...
struct	RectifierNeuron
	Rectifier Neuron f(x) = max(0,x) More...
struct	ReferenceVectorAdaptation
	Reference vector adaptation for the RVEA algorithm. More...
struct	ReferenceVectorGuidedSelection
	Implements the reference vector selection for the RVEA algorithm. More...
class	RegularizationNetworkTrainer
	Training of a regularization network. More...
class	RegularizedKernelMatrix
	Kernel Gram matrix with modified diagonal. More...
class	RemoveBudgetMaintenanceStrategy
	Budget maintenance strategy that removes a vector. More...
struct	ResultSet
class	RFClassifier
	Random Forest Classifier. More...
class	RFTrainer
	Random Forest. More...
class	RFTrainer< RealVector >
class	RFTrainer< unsigned int >
class	ROC
	ROC-Curve - false negatives over false positives. More...
struct	Rosenbrock
	Generalized Rosenbrock benchmark function. More...
struct	RotatedObjectiveFunction
	Rotates an objective function using a randomly initialized rotation. More...
struct	RouletteWheelSelection
	Fitness-proportional selection operator. More...
class	RpropMinus
	This class offers methods for the usage of the Resilient-Backpropagation-algorithm without weight-backtracking. More...
class	RpropPlus
	This class offers methods for the usage of the Resilient-Backpropagation-algorithm with weight-backtracking. More...
class	RVEA
	Implements the RVEA algorithm. More...
class	ScaledKernel
	Scaled version of a kernel function. More...
struct	Schwefel
	Convex benchmark function. More...
class	ScopedHandle
class	Shape
	Represents the Shape of an input or output. More...
class	Shark
	Allows for querying compile settings at runtime. Provides the current command line arguments to the rest of the library. More...
class	Shifter
	Shifter problem. More...
class	SimpleNearestNeighbors
	Brute force optimized nearest neighbor implementation. More...
class	SimplexDownhill
	Simplex Downhill Method. More...
struct	SimulatedBinaryCrossover
	Simulated binary crossover operator. More...
class	SingleChainApproximator
	Approximates the gradient by taking samples from a single Markov chain. More...
struct	SingleObjectiveResultSet
	Result set for single objective algorithm. More...
class	SinglePole
class	SMSEMOA
	Implements the SMS-EMOA. More...
class	SoftClusteringModel
	Model for "soft" clustering. More...
struct	SoftmaxNeuron
struct	Sphere
	Convex quadratic benchmark function. More...
class	SquaredEpsilonHingeLoss
	Hinge-loss for large margin regression using th squared two-norm. More...
class	SquaredHingeCSvmTrainer
class	SquaredHingeLinearCSvmTrainer
class	SquaredHingeLoss
	Squared Hinge-loss for large margin classification. More...
class	SquaredLoss
	squared loss for regression and classification More...
class	SquaredLoss< OutputType, unsigned int >
class	SquaredLoss< Sequence, Sequence >
struct	State
	Represents the State of an Object. More...
class	SteepestDescent
	Standard steepest descent. More...
class	SubrangeKernel
	Weighted sum of kernel functions. More...
class	SvmLogisticInterpretation
	Maximum-likelihood model selection score for binary support vector machines. More...
class	SvmProblem
class	SvmShrinkingProblem
struct	TanhNeuron
	Neuron which computes the hyperbolic tangenst with range [-1,1]. More...
class	TemperedMarkovChain
class	Timer
	Timer abstraction with microsecond resolution. More...
struct	TournamentSelection
	Tournament selection operator. More...
class	TrainingError
	This stopping criterion tracks the improvement of the error function of the training error over an interval of iterations. More...
class	TrainingProgress
	This stopping criterion tracks the improvement of the training error over an interval of iterations. More...
struct	TransformedData
class	TreeConstruction
	Stopping criteria for tree construction. More...
class	TreeNearestNeighbors
	Nearest Neighbors implementation using binary trees. More...
class	TruncatedExponentialLayer
	A layer of truncated exponential neurons. More...
class	TrustRegionNewton
	Simple Trust-Region method based on the full Hessian matrix. More...
class	TukeyBiweightLoss
	Tukey's Biweight-loss for robust regression. More...
class	TwoNormRegularizer
	Two-norm of the input as an objective function. More...
class	TwoPointStepSizeAdaptation
	Step size adaptation based on the success of the new population compared to the old. More...
struct	TwoStateSpace
	The TwoStateSpace is a discrete Space with only two values, for example {0,1} or {-1,1}. More...
class	TypedFeatureNotAvailableException
	Exception indicating the attempt to use a feature which is not supported. More...
class	TypedFlags
	Flexible and extensible mechanisms for holding flags. More...
class	UniformCrossover
	Uniform crossover of arbitrary individuals. More...
struct	UniformRankingSelection
	Selects individuals from the range of individual and offspring individuals. More...
class	UnlabeledData
	Data set for unsupervised learning. More...
struct	ValidatedSingleObjectiveResultSet
	Result set for validated points. More...
class	ValidatedStoppingCriterion
	Given the current Result set of the optimizer, calculates the validation error using a validation function and hands the results over to the underlying stopping criterion. More...
class	VariationalAutoencoderError
	Computes the variational autoencoder error function. More...
class	VDCMA
class	Wave
	Noisy sinc function: y = sin(x) / x + noise. More...
struct	WeightedDataBatch
struct	WeightedDataPair
	Input-Label pair of data. More...
class	WeightedLabeledData
	Weighted data set for supervised learning. More...
class	WeightedSumKernel
	Weighted sum of kernel functions. More...
class	WeightedUnlabeledData
	Weighted data set for unsupervised learning. More...
struct	WS2MaximumGradientCriterion
	Working set selection by maximization of the projected gradient. More...
struct	ZDT1
	Multi-objective optimization benchmark function ZDT1. More...
struct	ZDT2
	Multi-objective optimization benchmark function ZDT2. More...
struct	ZDT3
	Multi-objective optimization benchmark function ZDT3. More...
struct	ZDT4
	Multi-objective optimization benchmark function ZDT4. More...
struct	ZDT6
	Multi-objective optimization benchmark function ZDT6. More...
class	ZeroOneLoss
	0-1-loss for classification. More...
class	ZeroOneLoss< unsigned int, RealVector >
	0-1-loss for classification. More...

Typedefs
typedef IndicatorBasedMOCMA< HypervolumeIndicator >	MOCMA
typedef IndicatorBasedMOCMA< AdditiveEpsilonIndicator >	EpsilonMOCMA
typedef std::pair< double, RealVector >	Preference
	A preferred region in a lattice-sampled unit sphere. More...
typedef IndicatorBasedRealCodedNSGAII< HypervolumeIndicator >	RealCodedNSGAII
typedef IndicatorBasedRealCodedNSGAII< AdditiveEpsilonIndicator >	EpsRealCodedNSGAII
typedef IndicatorBasedRealCodedNSGAII< CrowdingDistance >	CrowdingRealCodedNSGAII
typedef IndicatorBasedSteadyStateMOCMA< HypervolumeIndicator >	SteadyStateMOCMA
typedef IndicatorBasedSteadyStateMOCMA< AdditiveEpsilonIndicator >	EpsilonSteadyStateMOCMA
typedef IRpropPlus	Rprop99
typedef IRpropMinus	Rprop99d
typedef RpropPlus	Rprop93
typedef RpropMinus	Rprop94
typedef boost::archive::polymorphic_iarchive	InArchive
	Type of an archive to read from. More...
typedef boost::archive::polymorphic_text_iarchive	TextInArchive
typedef boost::archive::polymorphic_oarchive	OutArchive
	Type of an archive to write to. More...
typedef boost::archive::polymorphic_text_oarchive	TextOutArchive
typedef std::pair< std::vector< std::size_t >, std::vector< std::size_t > >	RecreationIndices
	auxiliary typedef for createCVSameSizeBalanced and createCVFullyIndexed, stores location index in the first and partition index in the second More...
typedef LabeledData< RealVector, unsigned int >	ClassificationDataset
	specialized template for classification with unsigned int labels More...
typedef LabeledData< RealVector, RealVector >	RegressionDataset
	specialized template for regression with RealVector labels More...
typedef LabeledData< CompressedRealVector, unsigned int >	CompressedClassificationDataset
	specialized template for classification with unsigned int labels and sparse data More...
typedef blas::permutation_matrix	PermutationMatrix
typedef std::deque< RealVector >	Sequence
	Type of Data sequences. More...
typedef ARDKernelUnconstrained	DenseARDKernel
typedef ARDKernelUnconstrained< CompressedRealVector >	CompressedARDKernel
typedef GaussianRbfKernel	DenseRbfKernel
typedef GaussianRbfKernel< CompressedRealVector >	CompressedRbfKernel
typedef LinearKernel	DenseLinearKernel
typedef LinearKernel< CompressedRealVector >	CompressedLinearKernel
typedef ModelKernel< RealVector >	DenseModelKernel
typedef ModelKernel< CompressedRealVector >	CompressedModelKernel
typedef MonomialKernel	DenseMonomialKernel
typedef MonomialKernel< CompressedRealVector >	CompressedMonomialKernel
typedef NormalizedKernel	DenseNormalizedKernel
typedef NormalizedKernel< CompressedRealVector >	CompressedNormalizedKernel
typedef PolynomialKernel	DensePolynomialKernel
typedef PolynomialKernel< CompressedRealVector >	CompressedPolynomialKernel
typedef ScaledKernel	DenseScaledKernel
typedef ScaledKernel< CompressedRealVector >	CompressedScaledKernel
typedef SubrangeKernel< RealVector >	DenseSubrangeKernel
typedef SubrangeKernel< CompressedRealVector >	CompressesSubrangeKernel
typedef WeightedSumKernel	DenseWeightedSumKernel
typedef WeightedSumKernel< CompressedRealVector >	CompressedWeightedSumKernel
typedef AbstractObjectiveFunction< RealVector, double >	SingleObjectiveFunction
typedef AbstractObjectiveFunction< RealVector, RealVector >	MultiObjectiveFunction
typedef RBM< BinaryLayer, BinaryLayer, random::rng_type >	BinaryRBM
typedef GibbsOperator< BinaryRBM >	BinaryGibbsOperator
typedef MarkovChain< BinaryGibbsOperator >	BinaryGibbsChain
typedef TemperedMarkovChain< BinaryGibbsOperator >	BinaryPTChain
typedef MultiChainApproximator< BinaryGibbsChain >	BinaryPCD
typedef ContrastiveDivergence< BinaryGibbsOperator >	BinaryCD
typedef SingleChainApproximator< BinaryPTChain >	BinaryParallelTempering
typedef RBM< BipolarLayer, BipolarLayer, random::rng_type >	BipolarRBM
typedef GibbsOperator< BipolarRBM >	BipolarGibbsOperator
typedef MarkovChain< BipolarGibbsOperator >	BipolarGibbsChain
typedef TemperedMarkovChain< BipolarGibbsOperator >	BipolarPTChain
typedef MultiChainApproximator< BipolarGibbsChain >	BipolarPCD
typedef ContrastiveDivergence< BipolarGibbsOperator >	BipolarCD
typedef SingleChainApproximator< BipolarPTChain >	BipolarParallelTempering
typedef RBM< GaussianLayer, BinaryLayer, random::rng_type >	GaussianBinaryRBM
typedef GibbsOperator< GaussianBinaryRBM >	GaussianBinaryGibbsOperator
typedef MarkovChain< GaussianBinaryGibbsOperator >	GaussianBinaryGibbsChain
typedef TemperedMarkovChain< GaussianBinaryGibbsOperator >	GaussianBinaryPTChain
typedef MultiChainApproximator< GaussianBinaryGibbsChain >	GaussianBinaryPCD
typedef ContrastiveDivergence< GaussianBinaryGibbsOperator >	GaussianBinaryCD
typedef SingleChainApproximator< GaussianBinaryPTChain >	GaussianBinaryParallelTempering
typedef TwoStateSpace< 0, 1 >	BinarySpace
typedef TwoStateSpace<-1, 1 >	SymmetricBinarySpace
typedef RBM< TruncExpBinaryEnergy, random::rng_type >	TruncExpBinaryRBM
typedef GibbsOperator< TruncExpBinaryRBM >	TruncExpBinaryGibbsOperator
typedef MarkovChain< TruncExpBinaryGibbsOperator >	TruncExpBinaryGibbsChain
typedef TemperedMarkovChain< TruncExpBinaryGibbsOperator >	TruncExpBinaryPTChain
typedef MultiChainApproximator< TruncExpBinaryGibbsChain >	TruncExpBinaryPCD
typedef ContrastiveDivergence< TruncExpBinaryGibbsOperator >	TruncExpBinaryCD
typedef SingleChainApproximator< TruncExpBinaryPTChain >	TruncExpBinaryParallelTempering
typedef std::vector< shark::RealVector >	FrontType

Enumerations
enum	DominanceRelation { INCOMPARABLE = 0, LHS_DOMINATES_RHS = 1, RHS_DOMINATES_LHS = 2, EQUIVALENT = 3 }
	Result of comparing two objective vectors w.r.t. Pareto dominance. More...
enum	AlphaStatus { AlphaFree = 0, AlphaLowerBound = 1, AlphaUpperBound = 2, AlphaDeactivated = 3 }
enum	QpStopType { QpNone = 0, QpAccuracyReached = 1, QpMaxIterationsReached = 4, QpTimeout = 8 }
enum	McSvm {   McSvm::WW, McSvm::CS, McSvm::LLW, McSvm::ATM,   McSvm::ATS, McSvm::ADM, McSvm::OVA, McSvm::MMR,   McSvm::ReinforcedSvm }
enum	BuildType { RELEASE_BUILD_TYPE, DEBUG_BUILD_TYPE }
	Models the build type. More...
enum	Convolution { Convolution::Valid, Convolution::ZeroPad }
enum	PartitionEstimationAlgorithm {   AIS, AISMean, TwoSidedAISMean, AcceptanceRatio,   AcceptanceRatioMean }

Functions
SHARK_EXPORT_SYMBOL KernelExpansion< RealVector >	approximateKernelExpansion (random::rng_type &rng, KernelExpansion< RealVector > const &model, std::size_t k, double precision=1.e-8)
	Approximates a kernel expansion by a smaller one using an optimized basis. More...
template<class IndividualRange >
auto	penalizedFitness (IndividualRange &range) -> decltype(boost::adaptors::transform(range, detail::IndividualPenalizedFitnessFunctor()))
template<class IndividualRange >
auto	unpenalizedFitness (IndividualRange &range) -> decltype(boost::adaptors::transform(range, detail::IndividualUnpenalizedFitnessFunctor()))
template<class IndividualRange >
auto	ranks (IndividualRange &range) -> decltype(boost::adaptors::transform(range, detail::IndividualRankFunctor()))
template<class IndividualRange >
auto	searchPoint (IndividualRange &range) -> decltype(boost::adaptors::transform(range, detail::IndividualSearchPointFunctor()))
template<class PointRange , class RankRange >
void	dcNonDominatedSort (PointRange const &points, RankRange &ranks)
template<class PointRange , class RankRange >
void	fastNonDominatedSort (PointRange const &points, RankRange &ranks)
	Implements the well-known non-dominated sorting algorithm. More...
template<class PointRange , class RankRange >
void	nonDominatedSort (PointRange const &points, RankRange &ranks)
	Frontend for non-dominated sorting algorithms. More...
template<class PointRange , class RankRange >
void	nonDominatedSort (PointRange const &points, RankRange const &ranks)
template<class VectorTypeA , class VectorTypeB >
DominanceRelation	dominance (VectorTypeA const &lhs, VectorTypeB const &rhs)
	Pareto dominance relation for two objective vectors. More...
template<typename Matrix , typename randomType = shark::random::rng_type>
Matrix	sampleLatticeUniformly (randomType &rng, Matrix const &matrix, std::size_t const n, bool const keep_corners=true)
RealMatrix	preferenceAdjustedUnitVectors (std::size_t const n, std::size_t const sum, std::vector< Preference > const &preferences)
	Return a set of evenly spaced n-dimensional points on the unit sphere clustered around the specified preference points. More...
RealMatrix	preferenceAdjustedWeightVectors (std::size_t const n, std::size_t const sum, std::vector< Preference > const &preferences)
	Return a set of of evenly spaced n-dimensional points on the "unit simplex" clustered around the specified preference points. More...
std::size_t	computeOptimalLatticeTicks (std::size_t const n, std::size_t const target_count)
	Computes the number of Ticks for a grid of a certain size. More...
RealMatrix	weightLattice (std::size_t const n, std::size_t const sum)
	Returns a set of evenly spaced n-dimensional points on the "unit simplex". More...
RealMatrix	unitVectorsOnLattice (std::size_t const n, std::size_t const sum)
	Return a set of evenly spaced n-dimensional points on the unit sphere. More...
template<typename Matrix >
UIntMatrix	computeClosestNeighbourIndicesOnLattice (Matrix const &m, std::size_t const n)
double	tchebycheffScalarizer (RealVector const &fitness, RealVector const &weights, RealVector const &optimalPointFitness)
SHARK_EXPORT_SYMBOL std::size_t	kMeans (Data< RealVector > const &data, std::size_t k, Centroids &centroids, std::size_t maxIterations=0)
	The k-means clustering algorithm. More...
SHARK_EXPORT_SYMBOL std::size_t	kMeans (Data< RealVector > const &data, RBFLayer &model, std::size_t maxIterations=0)
	The k-means clustering algorithm for initializing an RBF Layer. More...
template<class InputType >
KernelExpansion< InputType >	kMeans (Data< InputType > const &dataset, std::size_t k, AbstractKernelFunction< InputType > &kernel, std::size_t maxIterations=0)
	The kernel k-means clustering algorithm. More...
template<class T >
T	maxExpInput ()
	Maximum allowed input value for exp. More...
template<class T >
T	minExpInput ()
	Minimum value for exp(x) allowed so that it is not 0. More...
template<class T >
boost::enable_if< std::is_arithmetic< T >, T >::type	sqr (const T &x)
	Calculates x^2. More...
template<class T >
T	cube (const T &x)
	Calculates x^3. More...
template<class T >
boost::enable_if< std::is_arithmetic< T >, T >::type	sigmoid (T x)
	Logistic function/logistic function. More...
template<class T >
T	safeExp (T x)
	Thresholded exp function, over- and underflow safe. More...
template<class T >
T	safeLog (T x)
	Thresholded log function, over- and underflow safe. More...
template<class T >
boost::enable_if< std::is_arithmetic< T >, T >::type	softPlus (T x)
	Numerically stable version of the function log(1+exp(x)). More...
double	softPlus (double x)
	Numerically stable version of the function log(1+exp(x)). calculated with float precision to save some time. More...
template<class T >
T	copySign (T x, T y)
template<typename T , typename U >
ResultSet< T, U >	makeResultSet (T const &t, U const &u)
	Generates a typed solution given the search point and the corresponding objective function value. More...
template<class SearchPoint , class Result >
std::ostream &	operator<< (std::ostream &out, ResultSet< SearchPoint, Result > const &solution)
template<class SearchPoint >
std::ostream &	operator<< (std::ostream &out, ValidatedSingleObjectiveResultSet< SearchPoint > const &solution)
bool	operator== (Shape const &shape1, Shape const &shape2)
bool	operator!= (Shape const &shape1, Shape const &shape2)
template<class E , class T >
std::basic_ostream< E, T > &	operator<< (std::basic_ostream< E, T > &os, Shape const &shape)
template<class Iterator , class Rng >
void	shuffle (Iterator begin, Iterator end, Rng &rng)
	random_shuffle algorithm which stops after acquiring the random subsequence for [begin,middle) More...
template<class RandomAccessIterator , class Rng >
void	partial_shuffle (RandomAccessIterator begin, RandomAccessIterator middle, RandomAccessIterator end, Rng &rng)
	random_shuffle algorithm which stops after acquiring the random subsequence for [begin,middle) More...
template<class RandomAccessIterator >
void	partial_shuffle (RandomAccessIterator begin, RandomAccessIterator middle, RandomAccessIterator end)
	random_shuffle algorithm which stops after acquiring the random subsequence for [begin,middle) More...
template<class Range >
boost::range_iterator< Range >::type	median_element (Range &range)
	Returns the iterator to the median element. after this call, the range is partially ordered. More...
template<class Range >
boost::range_iterator< Range >::type	median_element (Range const &rangeAdaptor)
template<class Range >
boost::range_iterator< Range >::type	partitionEqually (Range &range)
	Partitions a range in two parts as equal in size as possible. More...
template<class Range >
boost::range_iterator< Range >::type	partitionEqually (Range const &rangeAdaptor)
	Partitions a range in two parts as equal in size as possible and returns it's result. More...
template<class K , class V >
void	swap (KeyValuePair< K, V > &pair1, KeyValuePair< K, V > &pair2)
	Swaps the contents of two instances of KeyValuePair. More...
template<class Key , class Value >
KeyValuePair< Key, Value >	makeKeyValuePair (Key const &key, Value const &value)
	Creates a KeyValuePair. More...
template<class T , class Range >
Batch< T >::type	createBatch (Range const &range)
	creates a batch from a range of inputs More...
template<class Range >
Batch< typename Range::value_type >::type	createBatch (Range const &range)
	creates a batch from a range of inputs More...
template<class T , class Iterator >
Batch< T >::type	createBatch (Iterator const &begin, Iterator const &end)
template<class BatchT >
auto	getBatchElement (BatchT &batch, std::size_t i) -> decltype(BatchTraits< BatchT >::type::get(std::declval< BatchT &>(), i))
template<class BatchT >
auto	getBatchElement (BatchT const &batch, std::size_t i) -> decltype(BatchTraits< BatchT >::type::get(std::declval< BatchT const &>(), i))
template<class BatchT >
std::size_t	batchSize (BatchT const &batch)
template<class BatchT >
auto	batchBegin (BatchT &batch) -> decltype(BatchTraits< BatchT >::type::begin(batch))
template<class BatchT >
auto	batchBegin (BatchT const &batch) -> decltype(BatchTraits< BatchT >::type::begin(batch))
template<class BatchT >
auto	batchEnd (BatchT &batch) -> decltype(BatchTraits< BatchT >::type::end(batch))
template<class BatchT >
auto	batchEnd (BatchT const &batch) -> decltype(BatchTraits< BatchT >::type::end(batch))
template<class S >
S &	fusionize (detail::FusionFacade< S > &facade)
template<class S >
S const &	fusionize (detail::FusionFacade< S > const &facade)
template<class S >
boost::disable_if< detail::isFusionFacade< S >, S & >::type	fusionize (S &facade)
template<class S >
boost::disable_if< detail::isFusionFacade< S >, S const &>::type	fusionize (S const &facade)
template<class Range >
Data< typename Range::value_type >	createDataFromRange (Range const &inputs, std::size_t maximumBatchSize=0)
	creates a data object from a range of elements More...
template<class Range >
UnlabeledData< typename boost::range_value< Range >::type >	createUnlabeledDataFromRange (Range const &inputs, std::size_t maximumBatchSize=0)
	creates a data object from a range of elements More...
template<class Range1 , class Range2 >
LabeledData< typename boost::range_value< Range1 >::type, typename boost::range_value< Range2 >::type >	createLabeledDataFromRange (Range1 const &inputs, Range2 const &labels, std::size_t maximumBatchSize=0)
	creates a labeled data object from two ranges, representing inputs and labels More...
template<class T , class U >
std::ostream &	operator<< (std::ostream &stream, const LabeledData< T, U > &d)
	brief Outstream of elements for labeled data. More...
unsigned int	numberOfClasses (Data< unsigned int > const &labels)
	Return the number of classes of a set of class labels with unsigned int label encoding. More...
std::vector< std::size_t >	classSizes (Data< unsigned int > const &labels)
	Returns the number of members of each class in the dataset. More...
template<class InputType >
std::size_t	dataDimension (Data< InputType > const &dataset)
	Return the dimensionality of a dataset. More...
template<class InputType , class LabelType >
std::size_t	inputDimension (LabeledData< InputType, LabelType > const &dataset)
	Return the input dimensionality of a labeled dataset. More...
template<class InputType , class LabelType >
std::size_t	labelDimension (LabeledData< InputType, LabelType > const &dataset)
	Return the label/output dimensionality of a labeled dataset. More...
template<class InputType >
std::size_t	numberOfClasses (LabeledData< InputType, unsigned int > const &dataset)
	Return the number of classes (highest label value +1) of a classification dataset with unsigned int label encoding. More...
template<class InputType , class LabelType >
std::vector< std::size_t >	classSizes (LabeledData< InputType, LabelType > const &dataset)
	Returns the number of members of each class in the dataset. More...
template<class T , class Functor >
boost::lazy_disable_if< CanBeCalled< Functor, typename Data< T >::batch_type >, TransformedData< Functor, T >>::type	transform (Data< T > const &data, Functor f)
	Transforms a dataset using a Functor f and returns the transformed result. More...
template<class T , class Functor >
boost::lazy_enable_if< CanBeCalled< Functor, typename Data< T >::batch_type >, TransformedData< Functor, T >>::type	transform (Data< T > const &data, Functor const &f)
	Transforms a dataset using a Functor f and returns the transformed result. More...
template<class I , class L , class Functor >
LabeledData< typename detail::TransformedDataElement< Functor, I >::type, L >	transformInputs (LabeledData< I, L > const &data, Functor const &f)
	Transforms the inputs of a dataset and return the transformed result. More...
template<class I , class L , class Functor >
LabeledData< I, typename detail::TransformedDataElement< Functor, L >::type >	transformLabels (LabeledData< I, L > const &data, Functor const &f)
	Transforms the labels of a dataset and returns the transformed result. More...
template<class T , class FeatureSet >
Data< blas::vector< T > >	selectFeatures (Data< blas::vector< T > > const &data, FeatureSet const &features)
	Creates a copy of a dataset selecting only a certain set of features. More...
template<class T , class FeatureSet >
LabeledData< RealVector, T >	selectInputFeatures (LabeledData< RealVector, T > const &data, FeatureSet const &features)
template<class DatasetT >
DatasetT	splitAtElement (DatasetT &data, std::size_t elementIndex)
	Removes the last part of a given dataset and returns a new split containing the removed elements. More...
template<class I >
void	repartitionByClass (LabeledData< I, unsigned int > &data, std::size_t batchSize=LabeledData< I, unsigned int >::DefaultBatchSize)
	reorders the dataset such, that points are grouped by labels More...
template<class I >
LabeledData< I, unsigned int >	binarySubProblem (LabeledData< I, unsigned int >const &data, unsigned int zeroClass, unsigned int oneClass)
template<class I >
LabeledData< I, unsigned int >	oneVersusRestProblem (LabeledData< I, unsigned int >const &data, unsigned int oneClass)
	Construct a binary (two-class) one-versus-rest problem from a multi-class problem. More...
template<typename RowType >
RowType	getColumn (Data< RowType > const &data, std::size_t columnID)
template<typename RowType >
void	setColumn (Data< RowType > &data, std::size_t columnID, RowType newColumn)
template<class DatasetType , class IndexRange >
DataView< DatasetType >	subset (DataView< DatasetType > const &view, IndexRange const &indizes)
	creates a subset of a DataView with elements indexed by indices More...
template<class DatasetType >
DataView< DatasetType >	randomSubset (DataView< DatasetType > const &view, std::size_t size)
	creates a random subset of a DataView with given size More...
template<class DatasetType , class IndexRange >
DataView< DatasetType >::batch_type	subBatch (DataView< DatasetType > const &view, IndexRange const &indizes)
	Creates a batch given a set of indices. More...
template<class DatasetType >
DataView< DatasetType >::batch_type	randomSubBatch (DataView< DatasetType > const &view, std::size_t size)
	Creates a random batch of a given size. More...
template<class DatasetType >
DataView< DatasetType >	toView (DatasetType &set)
	Creates a View from a dataset. More...
template<class T >
DataView< T >::dataset_type	toDataset (DataView< T > const &view, std::size_t batchSize=DataView< T >::dataset_type::DefaultBatchSize)
	Creates a new dataset from a View. More...
template<class DatasetType >
std::size_t	numberOfClasses (DataView< DatasetType > const &view)
template<class DatasetType >
std::size_t	inputDimension (DataView< DatasetType > const &view)
	Return the input dimensionality of the labeled dataset represented by the view. More...
template<class DatasetType >
std::size_t	labelDimension (DataView< DatasetType > const &view)
	Return the label dimensionality of the labeled dataset represented by the view. More...
template<class DatasetType >
std::size_t	dataDimension (DataView< DatasetType > const &view)
	Return the dimensionality of the dataset represented by the view. More...
template<typename VectorType >
void	importHDF5 (Data< VectorType > &data, const std::string &fileName, const std::string &datasetName)
	Import data from a HDF5 file. More...
template<typename VectorType , typename LabelType >
void	importHDF5 (LabeledData< VectorType, LabelType > &labeledData, const std::string &fileName, const std::string &data, const std::string &label)
	Import data to a LabeledData object from a HDF5 file. More...
template<typename VectorType >
void	importHDF5 (Data< VectorType > &data, const std::string &fileName, const std::vector< std::string > &cscDatasetName)
	Import data from HDF5 dataset of compressed sparse column format. More...
template<typename VectorType , typename LabelType >
void	importHDF5 (LabeledData< VectorType, LabelType > &labeledData, const std::string &fileName, const std::vector< std::string > &cscDatasetName, const std::string &label)
	Import data from HDF5 dataset of compressed sparse column format. More...
template<class T >
void	importPGM (std::string const &fileName, T &data, std::size_t &sx, std::size_t &sy)
	Import a PGM image from file. More...
template<class T >
void	exportPGM (std::string const &fileName, T const &data, std::size_t sx, std::size_t sy, bool normalize=false)
	Export a PGM image to file. More...
void	exportFiltersToPGMGrid (std::string const &basename, RealMatrix const &filters, std::size_t width, std::size_t height)
	Exports a set of filters as a grid image. More...
void	exportFiltersToPGMGrid (std::string const &basename, Data< RealVector > const &filters, std::size_t width, std::size_t height)
	Exports a set of filters as a grid image. More...
template<class T >
void	importPGMSet (std::string const &p, Data< T > &set)
	Import PGM images scanning a directory recursively. More...
template<class Vec1T , class Vec2T , class Vec3T , class Device >
void	meanvar (Data< Vec1T > const &data, blas::vector_container< Vec2T, Device > &mean, blas::vector_container< Vec3T, Device > &variance)
	Calculates the mean and variance values of the input data. More...
template<class Vec1T , class Vec2T , class MatT , class Device >
void	meanvar (Data< Vec1T > const &data, blas::vector_container< Vec2T, Device > &mean, blas::matrix_container< MatT, Device > &variance)
	Calculates the mean, variance and covariance values of the input data. More...
template<class VectorType >
VectorType	mean (Data< VectorType > const &data)
	Calculates the mean vector of the input vectors. More...
template<class VectorType >
VectorType	mean (UnlabeledData< VectorType > const &data)
template<class VectorType >
VectorType	variance (Data< VectorType > const &data)
	Calculates the variance vector of the input vectors. More...
template<class VectorType >
blas::matrix< typename VectorType::value_type >	covariance (Data< VectorType > const &data)
	Calculates the covariance matrix of the data vectors. More...
template<class D1 , class W1 , class D2 , class W2 >
void	swap (WeightedDataPair< D1, W1 > &&p1, WeightedDataPair< D2, W2 > &&p2)
template<class D1 , class W1 , class D2 , class W2 >
void	swap (WeightedDataBatch< D1, W1 > &p1, WeightedDataBatch< D2, W2 > &p2)
template<class T >
std::ostream &	operator<< (std::ostream &stream, const WeightedUnlabeledData< T > &d)
	brief Outstream of elements for weighted data. More...
template<class DataRange , class WeightRange >
boost::disable_if< boost::is_arithmetic< WeightRange >, WeightedUnlabeledData< typename boost::range_value< DataRange >::type > >::type	createUnlabeledDataFromRange (DataRange const &data, WeightRange const &weights, std::size_t batchSize=0)
	creates a weighted unweighted data object from two ranges, representing data and weights More...
template<class T , class U >
std::ostream &	operator<< (std::ostream &stream, const WeightedLabeledData< T, U > &d)
	brief Outstream of elements for weighted labeled data. More...
std::size_t	numberOfClasses (WeightedUnlabeledData< unsigned int > const &labels)
std::vector< std::size_t >	classSizes (WeightedUnlabeledData< unsigned int > const &labels)
	Returns the number of members of each class in the dataset. More...
template<class InputType >
std::size_t	dataDimension (WeightedUnlabeledData< InputType > const &dataset)
	Return the dimnsionality of points of a weighted dataset. More...
template<class InputType , class LabelType >
std::size_t	inputDimension (WeightedLabeledData< InputType, LabelType > const &dataset)
	Return the input dimensionality of a weighted labeled dataset. More...
template<class InputType , class LabelType >
std::size_t	labelDimension (WeightedLabeledData< InputType, LabelType > const &dataset)
	Return the label/output dimensionality of a labeled dataset. More...
template<class InputType >
std::size_t	numberOfClasses (WeightedLabeledData< InputType, unsigned int > const &dataset)
	Return the number of classes (highest label value +1) of a classification dataset with unsigned int label encoding. More...
template<class InputType , class LabelType >
std::vector< std::size_t >	classSizes (WeightedLabeledData< InputType, LabelType > const &dataset)
	Returns the number of members of each class in the dataset. More...
template<class InputType >
double	sumOfWeights (WeightedUnlabeledData< InputType > const &dataset)
	Returns the total sum of weights. More...
template<class InputType , class LabelType >
double	sumOfWeights (WeightedLabeledData< InputType, LabelType > const &dataset)
	Returns the total sum of weights. More...
template<class InputType >
RealVector	classWeight (WeightedLabeledData< InputType, unsigned int > const &dataset)
	Computes the cumulative weight of every class. More...
template<class InputRange , class LabelRange , class WeightRange >
boost::disable_if< boost::is_arithmetic< WeightRange >, WeightedLabeledData< typename boost::range_value< InputRange >::type, typename boost::range_value< LabelRange >::type >>::type	createLabeledDataFromRange (InputRange const &inputs, LabelRange const &labels, WeightRange const &weights, std::size_t batchSize=0)
	creates a weighted unweighted data object from two ranges, representing data and weights More...
template<class InputType , class LabelType >
WeightedLabeledData< InputType, LabelType >	bootstrap (LabeledData< InputType, LabelType > const &dataset, std::size_t bootStrapSize=0)
	Creates a bootstrap partition of a labeled dataset and returns it using weighting. More...
template<class InputType >
WeightedUnlabeledData< InputType >	bootstrap (UnlabeledData< InputType > const &dataset, std::size_t bootStrapSize=0)
	Creates a bootstrap partition of an unlabeled dataset and returns it using weighting. More...
	SHARK_VECTOR_MATRIX_TYPEDEFS (long double, BigReal)
	SHARK_VECTOR_MATRIX_TYPEDEFS (bool, Bool)
template<class VectorType >
ConcatenatedModel< VectorType >	operator>> (AbstractModel< VectorType, VectorType, VectorType > &firstModel, AbstractModel< VectorType, VectorType, VectorType > &secondModel)
	Connects two AbstractModels so that the output of the first model is the input of the second. More...
template<class VectorType >
ConcatenatedModel< VectorType >	operator>> (ConcatenatedModel< VectorType > const &firstModel, AbstractModel< VectorType, VectorType, VectorType > &secondModel)
template<typename InputType , typename InputTypeT1 , typename InputTypeT2 >
double	evalSkipMissingFeatures (const AbstractKernelFunction< InputType > &kernelFunction, const InputTypeT1 &inputA, const InputTypeT2 &inputB)
template<typename InputType , typename InputTypeT1 , typename InputTypeT2 , typename InputTypeT3 >
double	evalSkipMissingFeatures (const AbstractKernelFunction< InputType > &kernelFunction, const InputTypeT1 &inputA, const InputTypeT2 &inputB, InputTypeT3 const &missingness)
template<class InputType , class M , class Device >
void	calculateRegularizedKernelMatrix (AbstractKernelFunction< InputType >const &kernel, Data< InputType > const &dataset, blas::matrix_expression< M, Device > &matrix, double regularizer=0)
	Calculates the regularized kernel gram matrix of the points stored inside a dataset. More...
template<class InputType , class M , class Device >
void	calculateMixedKernelMatrix (AbstractKernelFunction< InputType >const &kernel, Data< InputType > const &dataset1, Data< InputType > const &dataset2, blas::matrix_expression< M, Device > &matrix)
	Calculates the kernel gram matrix between two data sets. More...
template<class InputType >
RealMatrix	calculateRegularizedKernelMatrix (AbstractKernelFunction< InputType >const &kernel, Data< InputType > const &dataset, double regularizer=0)
	Calculates the regularized kernel gram matrix of the points stored inside a dataset. More...
template<class InputType >
RealMatrix	calculateMixedKernelMatrix (AbstractKernelFunction< InputType >const &kernel, Data< InputType > const &dataset1, Data< InputType > const &dataset2)
	Calculates the kernel gram matrix between two data sets. More...
template<class InputType , class WeightMatrix >
RealVector	calculateKernelMatrixParameterDerivative (AbstractKernelFunction< InputType > const &kernel, Data< InputType > const &dataset, WeightMatrix const &weights)
	Efficiently calculates the weighted derivative of a Kernel Gram Matrix w.r.t the Kernel Parameters. More...
template<class RBMType >
double	logPartitionFunction (RBMType const &rbm, double beta=1.0)
	Calculates the value of the partition function $Z$. More...
template<class RBMType >
double	negativeLogLikelihoodFromLogPartition (RBMType const &rbm, UnlabeledData< RealVector > const &inputs, double logPartition, double beta=1.0)
	Estimates the negative log-likelihood of a set of input vectors under the models distribution using the partition function. More...
template<class RBMType >
double	negativeLogLikelihood (RBMType const &rbm, UnlabeledData< RealVector > const &inputs, double beta=1.0)
	Estimates the negative log-likelihood of a set of input vectors under the models distribution. More...
double	estimateLogFreeEnergyFromEnergySamples (RealMatrix const &energyDiffUp, RealMatrix const &energyDiffDown, PartitionEstimationAlgorithm algorithm=AIS)
template<class RBMType >
double	estimateLogFreeEnergy (RBMType &rbm, Data< RealVector > const &initDataset, RealVector const &beta, std::size_t samples, PartitionEstimationAlgorithm algorithm=AcceptanceRatioMean, float burnInPercentage=0.1)
template<class RBMType >
double	annealedImportanceSampling (RBMType &rbm, RealVector const &beta, std::size_t samples)
template<class RBMType >
double	estimateLogFreeEnergy (RBMType &rbm, Data< RealVector > const &initDataset, std::size_t chains, std::size_t samples, PartitionEstimationAlgorithm algorithm=AIS, float burnInPercentage=0.1)
template<class RBMType >
double	annealedImportanceSampling (RBMType &rbm, std::size_t chains, std::size_t samples)
template<typename Stream >
FrontType	read_front (Stream &in, std::size_t noObjectives, const std::string &separator=" ", std::size_t headerLines=0)
template<class I , class L >
CVFolds< LabeledData< I, L > >	createCVIID (LabeledData< I, L > &set, std::size_t numberOfPartitions, std::size_t batchSize=Data< I >::DefaultBatchSize)
	Create a partition for cross validation. More...
template<class I , class L >
CVFolds< LabeledData< I, L > >	createCVSameSize (LabeledData< I, L > &set, std::size_t numberOfPartitions, std::size_t batchSize=LabeledData< I, L >::DefaultBatchSize)
	Create a partition for cross validation. More...
template<class I >
CVFolds< LabeledData< I, unsigned int > >	createCVSameSizeBalanced (LabeledData< I, unsigned int > &set, std::size_t numberOfPartitions, std::size_t batchSize=Data< I >::DefaultBatchSize, RecreationIndices *cv_indices=NULL)
	Create a partition for cross validation. More...
template<class I , class L >
CVFolds< LabeledData< I, L > >	createCVBatch (LabeledData< I, L > const &set, std::size_t numberOfPartitions)
	Create a partition for cross validation without changing the dataset. More...
template<class I , class L >
CVFolds< LabeledData< I, L > >	createCVIndexed (LabeledData< I, L > &set, std::size_t numberOfPartitions, std::vector< std::size_t > indices, std::size_t batchSize=Data< I >::DefaultBatchSize)
	Create a partition for cross validation from indices. More...
template<class I , class L >
CVFolds< LabeledData< I, L > >	createCVFullyIndexed (LabeledData< I, L > &set, std::size_t numberOfPartitions, RecreationIndices indices, std::size_t batchSize=Data< I >::DefaultBatchSize)
	Create a partition for cross validation from indices for both ordering and partitioning. More...
template<class T >
std::ostream &	operator<< (std::ostream &stream, const Data< T > &d)
	Outstream of elements. More...
SHARK_EXPORT_SYMBOL std::pair< std::string, std::string >	splitUrl (std::string const &url)
	Split a URL into its domain and resource parts. More...
SHARK_EXPORT_SYMBOL std::string	download (std::string const &url, unsigned short port=80)
	Download a document with the HTTP protocol. More...
template<class InputType , class LabelType >
void	downloadSparseData (LabeledData< InputType, LabelType > &dataset, std::string const &url, unsigned short port=80, unsigned int highestIndex=0, std::size_t batchSize=LabeledData< RealVector, unsigned int >::DefaultBatchSize)
	Download and import a sparse data (libSVM) file. More...
template<class InputType , class LabelType >
void	downloadFromMLData (LabeledData< InputType, LabelType > &dataset, std::string const &name, std::size_t batchSize=LabeledData< RealVector, unsigned int >::DefaultBatchSize)
	Download a data set from mldata.org. More...
template<class InputType >
void	downloadCsvData (LabeledData< InputType, unsigned int > &dataset, std::string const &url, LabelPosition lp, char separator=',', char comment='#', unsigned short port=80, std::size_t maximumBatchSize=LabeledData< RealVector, RealVector >::DefaultBatchSize)
	Download and import a dense data (CSV) file for classification. More...
template<class InputType >
void	downloadCsvData (LabeledData< InputType, RealVector > &dataset, std::string const &url, LabelPosition lp, std::size_t numberOfOutputs=1, char separator=',', char comment='#', unsigned short port=80, std::size_t maximumBatchSize=LabeledData< RealVector, RealVector >::DefaultBatchSize)
	Download and import a dense data (CSV) file for regression. More...
void	import_libsvm (LabeledData< RealVector, unsigned int > &dataset, std::istream &stream, unsigned int highestIndex=0, std::size_t batchSize=LabeledData< RealVector, unsigned int >::DefaultBatchSize)
	Import data from a LIBSVM file. More...
void	import_libsvm (LabeledData< CompressedRealVector, unsigned int > &dataset, std::istream &stream, unsigned int highestIndex=0, std::size_t batchSize=LabeledData< RealVector, unsigned int >::DefaultBatchSize)
	Import data from a LIBSVM file. More...
void	import_libsvm (LabeledData< RealVector, unsigned int > &dataset, std::string fn, unsigned int highestIndex=0, std::size_t batchSize=LabeledData< RealVector, unsigned int >::DefaultBatchSize)
	Import data from a LIBSVM file. More...
void	import_libsvm (LabeledData< CompressedRealVector, unsigned int > &dataset, std::string fn, unsigned int highestIndex=0, std::size_t batchSize=LabeledData< RealVector, unsigned int >::DefaultBatchSize)
	Import data from a LIBSVM file. More...
template<typename InputType >
void	export_libsvm (LabeledData< InputType, unsigned int > &dataset, const std::string &fn, bool dense=false, bool oneMinusOne=true, bool sortLabels=false, bool append=false)
	Export data to LIBSVM format. More...
SHARK_EXPORT_SYMBOL void	importSparseData (LabeledData< RealVector, unsigned int > &dataset, std::istream &stream, unsigned int highestIndex=0, std::size_t batchSize=LabeledData< RealVector, unsigned int >::DefaultBatchSize)
	Import classification data from a sparse data (libSVM) file. More...
SHARK_EXPORT_SYMBOL void	importSparseData (LabeledData< RealVector, RealVector > &dataset, std::istream &stream, unsigned int highestIndex=0, std::size_t batchSize=LabeledData< RealVector, RealVector >::DefaultBatchSize)
	Import regression data from a sparse data (libSVM) file. More...
SHARK_EXPORT_SYMBOL void	importSparseData (LabeledData< CompressedRealVector, unsigned int > &dataset, std::istream &stream, unsigned int highestIndex=0, std::size_t batchSize=LabeledData< RealVector, unsigned int >::DefaultBatchSize)
	Import classification data from a sparse data (libSVM) file. More...
SHARK_EXPORT_SYMBOL void	importSparseData (LabeledData< CompressedRealVector, RealVector > &dataset, std::istream &stream, unsigned int highestIndex=0, std::size_t batchSize=LabeledData< RealVector, RealVector >::DefaultBatchSize)
	Import regression data from a sparse data (libSVM) file. More...
SHARK_EXPORT_SYMBOL void	importSparseData (LabeledData< RealVector, unsigned int > &dataset, std::string fn, unsigned int highestIndex=0, std::size_t batchSize=LabeledData< RealVector, unsigned int >::DefaultBatchSize)
	Import classification data from a sparse data (libSVM) file. More...
SHARK_EXPORT_SYMBOL void	importSparseData (LabeledData< RealVector, RealVector > &dataset, std::string fn, unsigned int highestIndex=0, std::size_t batchSize=LabeledData< RealVector, RealVector >::DefaultBatchSize)
	Import regression data from a sparse data (libSVM) file. More...
SHARK_EXPORT_SYMBOL void	importSparseData (LabeledData< CompressedRealVector, unsigned int > &dataset, std::string fn, unsigned int highestIndex=0, std::size_t batchSize=LabeledData< RealVector, unsigned int >::DefaultBatchSize)
	Import classification data from a sparse data (libSVM) file. More...
SHARK_EXPORT_SYMBOL void	importSparseData (LabeledData< CompressedRealVector, RealVector > &dataset, std::string fn, unsigned int highestIndex=0, std::size_t batchSize=LabeledData< RealVector, RealVector >::DefaultBatchSize)
	Import regression data from a sparse data (libSVM) file. More...
template<typename InputType >
void	exportSparseData (LabeledData< InputType, unsigned int > const &dataset, std::ostream &stream, bool oneMinusOne=true, bool sortLabels=false)
	Export classification data to sparse data (libSVM) format. More...
template<typename InputType >
void	exportSparseData (LabeledData< InputType, unsigned int > const &dataset, const std::string &fn, bool oneMinusOne=true, bool sortLabels=false, bool append=false)
	Export classification data to sparse data (libSVM) format. More...
template<typename InputType >
void	exportSparseData (LabeledData< InputType, RealVector > const &dataset, std::ostream &stream)
	Export regression data to sparse data (libSVM) format. More...
template<typename InputType >
void	exportSparseData (LabeledData< InputType, RealVector > const &dataset, const std::string &fn, bool append=false)
	Export regression data to sparse data (libSVM) format. More...
template<class InputType , class OutputType >
void	initRandomNormal (AbstractModel< InputType, OutputType > &model, double s)
	Initialize model parameters normally distributed. More...
template<class InputType , class OutputType >
void	initRandomUniform (AbstractModel< InputType, OutputType > &model, double lower, double upper)
	Initialize model parameters uniformly at random. More...

Variables
static const double	SQRT_2_PI = boost::math::constants::root_two_pi<double>()
	Constant for sqrt( 2 * pi ). More...
enum	LabelPosition { FIRST_COLUMN, LAST_COLUMN }
	Position of the label in a CSV file. More...
SHARK_EXPORT_SYMBOL void	csvStringToData (Data< FloatVector > &data, std::string const &contents, char separator=',', char comment='#', std::size_t maximumBatchSize=Data< RealVector >::DefaultBatchSize)
	Import unlabeled vectors from a read-in character-separated value file. More...
SHARK_EXPORT_SYMBOL void	csvStringToData (Data< RealVector > &data, std::string const &contents, char separator=',', char comment='#', std::size_t maximumBatchSize=Data< RealVector >::DefaultBatchSize)
	Import unlabeled vectors from a read-in character-separated value file. More...
SHARK_EXPORT_SYMBOL void	csvStringToData (Data< unsigned int > &data, std::string const &contents, char separator=',', char comment='#', std::size_t maximumBatchSize=Data< unsigned int >::DefaultBatchSize)
	Import "csv" from string consisting only of a single unsigned int per row. More...
SHARK_EXPORT_SYMBOL void	csvStringToData (Data< int > &data, std::string const &contents, char separator=',', char comment='#', std::size_t maximumBatchSize=Data< int >::DefaultBatchSize)
	Import "csv" from string consisting only of a single int per row. More...
SHARK_EXPORT_SYMBOL void	csvStringToData (Data< float > &data, std::string const &contents, char separator=',', char comment='#', std::size_t maximumBatchSize=Data< double >::DefaultBatchSize)
	Import "csv" from string consisting only of a single double per row. More...
SHARK_EXPORT_SYMBOL void	csvStringToData (Data< double > &data, std::string const &contents, char separator=',', char comment='#', std::size_t maximumBatchSize=Data< double >::DefaultBatchSize)
	Import "csv" from string consisting only of a single double per row. More...
SHARK_EXPORT_SYMBOL void	csvStringToData (LabeledData< RealVector, unsigned int > &dataset, std::string const &contents, LabelPosition lp, char separator=',', char comment='#', std::size_t maximumBatchSize=LabeledData< RealVector, unsigned int >::DefaultBatchSize)
	Import labeled data from a character-separated value file. More...
SHARK_EXPORT_SYMBOL void	csvStringToData (LabeledData< FloatVector, unsigned int > &dataset, std::string const &contents, LabelPosition lp, char separator=',', char comment='#', std::size_t maximumBatchSize=LabeledData< RealVector, unsigned int >::DefaultBatchSize)
	Import labeled data from a character-separated value file. More...
SHARK_EXPORT_SYMBOL void	csvStringToData (LabeledData< RealVector, RealVector > &dataset, std::string const &contents, LabelPosition lp, std::size_t numberOfOutputs=1, char separator=',', char comment='#', std::size_t maximumBatchSize=LabeledData< RealVector, RealVector >::DefaultBatchSize)
	Import regression data from a read-in character-separated value file. More...
SHARK_EXPORT_SYMBOL void	csvStringToData (LabeledData< FloatVector, FloatVector > &dataset, std::string const &contents, LabelPosition lp, std::size_t numberOfOutputs=1, char separator=',', char comment='#', std::size_t maximumBatchSize=LabeledData< RealVector, RealVector >::DefaultBatchSize)
	Import regression data from a read-in character-separated value file. More...
template<class T >
void	importCSV (Data< T > &data, std::string fn, char separator=',', char comment='#', std::size_t maximumBatchSize=Data< T >::DefaultBatchSize, std::size_t titleLines=0)
	Import a Dataset from a csv file. More...
template<class T >
void	importCSV (LabeledData< blas::vector< T >, unsigned int > &data, std::string fn, LabelPosition lp, char separator=',', char comment='#', std::size_t maximumBatchSize=LabeledData< RealVector, unsigned int >::DefaultBatchSize)
	Import a labeled Dataset from a csv file. More...
template<class T >
void	importCSV (LabeledData< blas::vector< T >, blas::vector< T > > &data, std::string fn, LabelPosition lp, std::size_t numberOfOutputs=1, char separator=',', char comment='#', std::size_t maximumBatchSize=LabeledData< RealVector, RealVector >::DefaultBatchSize)
	Import a labeled Dataset from a csv file. More...
template<typename Type >
void	exportCSV (Data< Type > const &set, std::string fn, char separator=',', bool sci=true, unsigned int width=0)
	Format unlabeled data into a character-separated value file. More...
template<typename InputType , typename LabelType >
void	exportCSV (LabeledData< InputType, LabelType > const &dataset, std::string fn, LabelPosition lp, char separator=',', bool sci=true, unsigned int width=0)
	Format labeled data into a character-separated value file. More...
enum	KernelMatrixNormalizationType {   NONE, MULTIPLICATIVE_TRACE_ONE, MULTIPLICATIVE_TRACE_N, MULTIPLICATIVE_VARIANCE_ONE,   CENTER_ONLY, CENTER_AND_MULTIPLICATIVE_TRACE_ONE }
template<typename InputType , typename LabelType >
void	exportKernelMatrix (LabeledData< InputType, LabelType > const &dataset, AbstractKernelFunction< InputType > &kernel, std::ostream &out, KernelMatrixNormalizationType normalizer=NONE, bool scientific=false, unsigned int fieldwidth=0)
	Write a kernel Gram matrix to stream. More...
template<typename InputType , typename LabelType >
void	exportKernelMatrix (LabeledData< InputType, LabelType > const &dataset, AbstractKernelFunction< InputType > &kernel, std::string fn, KernelMatrixNormalizationType normalizer=NONE, bool sci=false, unsigned int width=0)
	Write a kernel Gram matrix to file. More...
template<typename InputType , typename LabelType >
void	export_kernel_matrix (LabeledData< InputType, LabelType > const &dataset, AbstractKernelFunction< InputType > &kernel, std::ostream &out, KernelMatrixNormalizationType normalizer=NONE, bool scientific=false, unsigned int fieldwidth=0)
template<typename InputType , typename LabelType >
void	export_kernel_matrix (LabeledData< InputType, LabelType > const &dataset, AbstractKernelFunction< InputType > &kernel, std::string fn, KernelMatrixNormalizationType normalizer=NONE, bool sci=false, unsigned int width=0)

Detailed Description

AbstractMultiObjectiveOptimizer.

Implements Block Gibbs Sampling.

Implements the Shifter benchmark problem.

Loads the MNIST benchmark problem.

Implements the DistantModes/ArtificialModes benchmark problem.

Implements the Bars & Stripes benchmark problem.

Implements the bipolar (-1,1) state neuron layer.

Typedefs for the Bipolar RBM.

Typedefs for the Binary-Binary RBM.

Calculate statistics given a range of values.

ROC.

Implements a multi-variate normal distribution with zero mean.

Implements a multinomial distribution.

Variational-autoencoder error function.

Maximum-likelihood model selection for binary support vector machines.

Regularizer.

Radius Margin Quotient for SVM model selection.

Negative Log Likelihood error function.

Evidence for model selection of a regularization network/Gaussian process.

Functions for measuring the area under the (ROC) curve.

Error measure for classication tasks, typically used for evaluation of results.

Implements Tukey's Biweight-loss function for robust regression.

Implements the Squared Error Loss function for regression.

Implements the Squared Hinge Loss function for maximum margin classification.

Implements the squard Hinge Loss function for maximum margin regression.

Implements the Huber loss function for robust regression.

Implements the Hinge Loss function for maximum margin classification.

Implements the Hinge Loss function for maximum margin regression.

Flexible error measure for classication tasks.

Error measure for classification tasks that can be used as the objective function for training.

super class of all loss functions

implements the absolute loss, which is the distance between labels and predictions

Leave-one-out error for C-SVMs.

Leave-one-out error.

Kernel Target Alignment - a measure of alignment of a kernel Gram matrix with labels.

Archive of evaluated points as an objective function wrapper.

error function for supervised learning

cross-validation error for selection of hyper-parameters

CombinedObjectiveFunction.

Multi-objective optimization benchmark function ZDT6.

Multi-objective optimization benchmark function ZDT4.

Multi-objective optimization benchmark function ZDT3.

Multi-objective optimization benchmark function ZDT2.

Multi-objective optimization benchmark function ZDT1.

Convex benchmark function.

Implements a wrapper over an m_objective function which just rotates its inputs.

Generalized Rosenbrock benchmark function.

Pole balancing simulation for double pole.

Pole balancing simulation for double poles.

Objective function for single and double poles with partial state information (non-Markovian task)

Objective function for single and double poles with full state information (Markovian task)

Multi-objective optimization benchmark function LZ9.

Multi-objective optimization benchmark function LZ8.

Multi-objective optimization benchmark function LZ7.

Multi-objective optimization benchmark function LZ6.

Multi-objective optimization benchmark function LZ5.

Multi-objective optimization benchmark function LZ4.

Multi-objective optimization benchmark function LZ3.

Multi-objective optimization benchmark function LZ2.

Multi-objective optimization benchmark function LZ1.

Multi-objective optimization benchmark function IHR 6.

Multi-objective optimization benchmark function IHR 4.

Multi-objective optimization benchmark function IHR 3.

Multi-objective optimization benchmark function IHR 2.

Multi-objective optimization benchmark function IHR 1.

Two-dimensional, real-valued Himmelblau function.

GSP benchmark function for multiobjective optimization.

Bi-objective real-valued benchmark function proposed by Fonseca and Flemming.

Multi-objective optimization benchmark function ELLI 2.

Multi-objective optimization benchmark function ELLI 1.

Objective function DTLZ7.

Objective function DTLZ6.

Objective function DTLZ5.

Objective function DTLZ4.

Objective function DTLZ3.

Objective function DTLZ2.

Objective function DTLZ1.

Multi-objective optimization benchmark function CIGTAB 2.

Multi-objective optimization benchmark function CIGTAB 1.

Convex quadratic benchmark function.

Convex quadratic benchmark function with single dominant axis.

AbstractObjectiveFunction.

cost function for quantitative judgement of deviations of predictions from target values

Base class for constraints.

Random Forest Classifier.

Tree for nearest neighbor search in data with low embedding dimension.

Tree for nearest neighbor search in kernel-induced feature spaces.

Tree for nearest neighbor search in low dimensions.

Cart Classifier.

Binary space-partitioning tree of data points.

Implements a radial basis function layer.

One-versus-one Classifier.

Model for scaling and translation of data vectors.

NEarest neighbor model for classification and regression.

Implements the Mean Model that can be used for ensemble classifiers.

Implements a Model using a linear function.

Weighted sum of m_base kernels.

Variant of WeightedSumKernel which works on subranges of Vector inputs.

A kernel function that wraps a member kernel and multiplies it by a scalar.

Product of kernel functions.

Polynomial kernel.

Applies a kernel to two pointsets and comptues the average response.

Normalization of a kernel function.

Special kernel classes for multi-task and transfer learning.

monomial (polynomial) kernel

Weighted sum of base kernels, each acting on a subset of features only.

A kernel expansion with support of missing features.

linear kernel (standard inner product)

Collection of functions dealing with typical tasks of kernels.

Affine linear kernel function expansion.

Radial Gaussian kernel.

Do special kernel evaluation by skipping missing features.

Kernel on a finite, discrete space.

Derivative of a C-SVM hypothesis w.r.t. its hyperparameters.

Gaussian automatic relevance detection (ARD) kernel.

abstract super class of all metrics

abstract super class of all kernel functions

Implements a model applying a convolution to an image.

concatenation of two models, with type erasure

Model for "soft" clustering.

Hierarchical Clustering.

Model for "hard" clustering.

Super class for clustering models.

Clusters defined by centroids.

Super class for clustering definitions.

Model for conversion of real valued output to class labels.

base class for all models, as well as a specialized differentiable model

Some operations for creating rotation matrices.

Kernel Gram matrix with modified diagonal.

Precomputed version of a matrix for quadratic programming.

Partly Precomputed version of a matrix for quadratic programming.

Modified Kernel Gram matrix.

Cache implementing an Least-Recently-Used Strategy.

Kernel Gram matrix.

Efficient special case if the kernel is gaussian and the inputs are sparse vectors.

Kernel matrix which supports kernel evaluations on data with missing features.

Kernel matrix for SVM ranking.

Efficient quadratic matrix cache.

Kernel matrix for SVM regression.

Entry Point for all Basic Linear Algebra(BLAS) in shark.

Weighted data sets for (un-)supervised learning.

some functions for vector valued statistics like mean, variance and covariance

Support for importing and exporting data from and to sparse data (libSVM) formatted data files.

Importing and exporting PGM images.

Deprecated import_libsvm and export_libsvm functions.

This will relabel a given dataset to have labels 0..N-1 (and vice versa)

Support for importing data from HDF5 file.

export precomputed kernel matrices (using libsvm format)

Support for downloading data sets from online sources.

Fast lookup for elements in constant datasets.

Data for (un-)supervised learning.

Learning problems given by analytic distributions.

Tools for cross-validation.

Support for importing and exporting data from and to character separated value (CSV) files.

Defines an batch adptor for structures.

Defines the Batch Interface for a type, e.g., for every type a container with optimal structure.

A scoped_ptr like container for C type handles.

Provides a pair of Key and Value, as well as functions working with them.

Small Iterator collection.

Small General algorithm collection.

Template class checking whether for a functor F and Argument U, F(U) can be called.

Traits which allow to define ProxyReferences for types.

Timer abstraction with microsecond resolution.

Class which externalizes the state of an Object.

Class Describing the Shape of an Input.

Result sets for algorithms.

Shark Random number generation.

Very basic math abstraction layer.

ISerializable interface.

IParameterizable interface.

INameable interface.

Flexible and extensible mechanisms for holding flags.

Exception.

Random Forest Trainer.

Trainer for a Regularization Network or a Gaussian Process.

Support Vector Machine Trainer for the ranking-SVM.

Perceptron.

Principal Component Analysis.

Model training by means of a general purpose optimization procedure.

Trainer for One-Class Support Vector Machines.

Determine the scaling factor of a ScaledKernel so that it has unit variance in feature space one on a given dataset.

Data normalization to zero mean, unit variance and zero covariance while keping the original coordinate system.

Data normalization to zero mean, unit variance and zero covariance.

Data normalization to zero mean and unit variance.

Data normalization to the unit interval.

Trainer for binary SVMs natively supporting missing features.

Logistic Regression.

Generic Stochastic Average Gradient Descent training for linear models.

Linear Regression.

LDA.

LASSO Regression.

Generic stochastic gradient descent training for kernel-based models.

KernelMeanClassifier.

FisherLDA.

Trainer for the Epsilon-Support Vector Machine for Regression.

Remove budget maintenance strategy.

Project budget maintenance strategy.

Merge budget maintenance strategy.

Budgeted stochastic gradient descent training for kernel-based models.

Abstract Budget maintenance strategy.

Abstract Trainer Interface for trainers that support weighting.

Abstract Trainer Interface.

Abstract Support Vector Machine Trainer, general and linear case.

Stopping Criterion which evaluates the validation error and hands the result over to another stopping criterion.

Stopping Criterion which stops, when the training error seems to converge.

Stopping Criterion which stops, when the trainign error seems to converge.

Stopping Criterion which stops after a fixed number of iterations.

Stopping criterion monitoring the quotient of generalization loss and training progress.

Stopping Criterion which stops, when the generalization of the solution gets worse.

Defines a base class for stopping criteria of optimization algorithms.

Quadratic programming for Support Vector Machines.

Special container for certain coefficients describing multi-class SVMs.

General and specialized quadratic program classes and a generic solver.

Quadratic programming problem for multi-class SVMs.

Quadratic programming solvers for linear multi-class SVM training without bias.

Quadratic programming m_problem for multi-class SVMs.

Quadratic programming solver linear SVM training without bias.

Quadratic program definitions.

Shrinking strategy based on box constraints.

Pegasos solvers for linear SVMs.

Efficient Nearest neighbor queries.

Efficient brute force implementation of nearest neighbors.

Interface for nearest Neighbor queries.

Jaakkola's heuristic and related quantities for Gaussian kernel selection.

Trust-Region Newton-Step Method.

SteepestDescent.

implements different versions of Resilient Backpropagation of error.

LineSearch.

LBFGS.

CG.

BFGS.

Adam.

Base class for Line Search Optimizer.

Implements the VD-CMA-ES Algorithm.

SteadyStateMOCMA.h.

Implements the SMS-EMOA.

Nelder-Mead Simplex Downhill Method.

Implements the RVEA algorithm.

RealCodedNSGAIIII.h.

IndicatorBasedRealCodedNSGAII.h.

Implements a two point step size adaptation rule based on a line-search type of approach.

Roulette-Wheel-Selection using uniform selection probability assignment.

Implements tournament selection.

Implements fitness proportional selection.

Implements the reference vector selection for RVEA.

Roulette-Wheel-Selection based on fitness-rank-based selection probability assignment.

Indicator-based selection strategy for multi-objective selection.

EP-Tournament selection operator.

Elitist Selection Operator suitable for (mu,lambda) and (mu+lambda) selection.

Implements the Tchebycheff scalarizer.

Reference vector adaptation for the RVEA algorithm.

Uniform crossover of arbitrary individuals.

Simulated binary crossover operator.

Implements one-point crossover operator.

Implements the tep size adaptation based on the success of the new population compared to the old.

Polynomial mutation operator.

Bit flip mutation operator.

Various functions for generating n-dimensional grids (simplex-lattices).

Algorithm selecting front based on their crowding distance.

Calculates the hypervolume covered by a front of non-dominated points.

Algorithm selecting points based on their crowding distance.

Calculates the additive approximation quality of a Pareto-front approximation.

Approximately determines the individual contributing the least hypervolume.

Implements the frontend for the HypervolumeContribution algorithms, including the approximations.

Implementation of the exact hypervolume calculation in m dimensions.

Implementation of the exact hypervolume calculation in 3 dimensions.

Implementation of the exact hypervolume calculation in 2 dimensions.

Implements the frontend for the HypervolumeCalculator algorithms, including the approximations.

Determine the volume of the union of objects by an FPRAS.

PenalizingEvaluator.

Implementation of the Pareto-Dominance relation.

Swapper method for non-dominated sorting.

Implements the fast non-dominated sort algorithm.

Implements a divide-and-conquer non-dominated sorting algorithm.

Implements the MOEA/D algorithm.

Implements the generational Multi-objective Covariance Matrix Adapation ES.

• GridSearch.h.

Implements the most recent version of the elitist CMA-ES.

Implements the Cross Entropy Algorithm.

Implements the CMSA.

Implements the most recent version of the non-elitist CMA-ES.

TypedIndividual.

CMAChromosomeof the CMA-ES.

The k-means clustering algorithm.

AbstractSingleObjectiveOptimizer.

AbstractOptimizer.

Author

T.Voss, T. Glasmachers, O.Krause

Date

2010-2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Hansen, N. The CMA Evolution Startegy: A Tutorial, June 28, 2011 and the eqation numbers refer to this publication (retrieved April 2014).

Author

Thomas Voss and Christian Igel

Date

April 2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

The algorithm is described in

H. G. Beyer, B. Sendhoff (2008). Covariance Matrix Adaptation Revisited: The CMSA Evolution Strategy In Proceedings of the Tenth International Conference on Parallel Problem Solving from Nature (PPSN X), pp. 123-132, LNCS, Springer-Verlag

Copyright (c) 1998-2008:

Institut für Neuroinformatik

Author

-

Date

-

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Christophe Thiery, Bruno Scherrer. Improvements on Learning Tetris with Cross Entropy. International Computer Games Association Journal, ICGA, 2009, 32. <inria-00418930>

Author

Jens Holm, Mathias Petræus and Mark Wulff

Date

January 2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

The algorithm is based on

C. Igel, T. Suttorp, and N. Hansen. A Computational Efficient Covariance Matrix Update and a (1+1)-CMA for Evolution Strategies. In Proceedings of the Genetic and Evolutionary Computation Conference (GECCO 2006), pp. 453-460, ACM Press, 2006

D. V. Arnold and N. Hansen: Active covariance matrix adaptation for the (1+1)-CMA-ES. In Proceedings of the Genetic and Evolutionary Computation Conference (GECCO 2010): pp 385-392, ACM Press 2010

Author

O. Krause T.Voss

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause

Date

2010

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Voss, T. Glasmachers, O.Krause

Date

2010-2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Thomas Voss and Christian Igel

Date

April 2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Voss

Date

2010

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Bjoern Bugge Grathwohl

Date

February 2017

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Voss, O. Krause

Date

2015

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers (based on old version by T. Voß)

Date

2011-2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Voss, O.Krause

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Voss

Date

2010-2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause

Date

2014-2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause, T. Glasmachers

Date

2014-2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

2016-2017

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Voss, O.Krause, T. Glasmachers

Date

2014-2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause

Date

2017

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Voss, O.Krause

Date

2010-2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause

Date

2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Voss, O.Krause

Date

2010-2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Bjørn Bugge Grathwohl

Date

February 2017

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Voss O.Krause

Date

2010-2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Voss

Date

2010-2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Bjoern Bugge Grathwohl

Date

March 2017

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

K. Miettinen, "Nonlinear Multiobjective Optimization", International Series in Operations Research and Management Science (12) DOI: 10.1007/978-1-4615-5563-6

Author

Bjørn Bugge Grathwohl

Date

February 2017

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

The algorithm is described in: James E. Baker. Adaptive Selection Methods for Genetic Algorithms. In John J. Grefenstette (ed.): Proceedings of the 1st International Conference on Genetic Algorithms (ICGA), pp. 101-111, Lawrence Erlbaum Associates, 1985

Author

T.Voss

Date

2010-2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

See http://en.wikipedia.org/wiki/Fitness_proportionate_selection

Author

T. Voss

Copyright (c) 1998-2008:

Institut für Neuroinformatik

Date

-

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

See http://en.wikipedia.org/wiki/Tournament_selection

Author

-

Date

-

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

2015

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

See Nicola Beume, Boris Naujoks, and Michael Emmerich. SMS-EMOA: Multiobjective selection based on dominated hypervolume. European Journal of Operational Research, 181(3):1653-1669, 2007.

Author

T.Voss

Date

2010

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Oswin Krause

Date

April 2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.Implements the VD-CMA-ES Algorithm

The VD-CMA-ES implements a restricted form of the CMA-ES where the covariance matrix is restriced to be (D+vv^T) where D is a diagonal matrix and v a single vector. Therefore this variant is capable of large-scale optimisation

For more reference, see the paper Akimoto, Y., A. Auger, and N. Hansen (2014). Comparison-Based Natural Gradient Optimization in High Dimension. To appear in Genetic and Evolutionary Computation Conference (GECCO 2014), Proceedings, ACM

The implementation differs from the paper to be closer to the reference implementation and to have better numerical accuracy.

Author

O. Krause

Date

2013

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause

Date

2017

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

The Broyden, Fletcher, Goldfarb, Shannon (BFGS) algorithm is a quasi-Newton method for unconstrained real-valued optimization.

Author

O. Krause

Date

2010

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Conjugate-gradient method for unconstraint optimization.

Author

O. Krause

Date

2010

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

The Limited-Memory Broyden, Fletcher, Goldfarb, Shannon (BFGS) algorithm is a quasi-Newton method for unconstrained real-valued optimization. See: http://en.wikipedia.org/wiki/LBFGS for details.

Author

S. Dahlgaard, O.Krause

Date

2013

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause, S. Dahlgaard

Date

2010-2017

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Oswin Krause

Date

2010

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause

Date

2015

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers, O. Krause, C. Igel

Date

2010

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause

Date

2012-2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers, O.Krause

Date

2017

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers, O.Krause

Date

2013

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

-

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

2007-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers, O.Krause

Date

2007-2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This file provides a number of classes representing hugh dense matrices all related to kernel Gram matices of possibly large datasets. These classes share a common interface for (a) providing a matrix entry, (b) swapping two variable indices, and (c) returning the matrix size.

This interface is required by the template class CachedMatrix, which provides a cache mechanism for restricted matrix rows, as it is used by various quadratic program solvers within the library. The PrecomputedMatrix provides a sometimes faster but more memory intensive alternative to CachedMatrix.

Author

T. Glasmachers

Date

2007-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This file provides: 1) the QpConfig class, which can configure and provide information about an SVM training procedure; 2) a super-class for general SVM trainers, namely the AbstractSvmTrainer; and 3) a streamlined variant thereof for purely linear SVMs, namely the AbstractLinearSvmTrainer. In general, the SvmTrainers hold as parameters all hyperparameters of the underlying SVM, which includes the kernel parameters for non-linear SVMs.

Author

T. Glasmachers

Date

-

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause, T.Glasmachers

Date

2010-2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This holds the interface for any budget maintenance strategy.

Author

Aydin Demircioglu

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This is an implementation of the BSGD algorithm, developed by

Wang, Crammer and Vucetic: Breaking the curse of kernelization: Budgeted stochastic gradient descent for large-scale SVM training, JMLR 2012. Basically this is pegasos, so something similar to a perceptron. The main difference is that we do restrict the sparsity of the weight vector to a (currently predefined) value. Therefore, whenever this sparsity is reached, we have to decide how to add a new vector to the model, without destroying this sparsity. Several methods have been proposed for this, Wang et al. main insight is that merging two budget vectors (i.e. two vectors in the model). If the first one is searched by norm of its alpha coefficient, the second one can be found by some optimization problem, yielding a roughly optimal pair. This pair can be merged and by doing so the budget has now space for a new vector. Such strategies are called budget maintenance strategies.

This implementation owes much to the 'reference' implementation

in the BudgetedSVM software.

Author

T. Glasmachers, Aydin Demircioglu

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This is an budget strategy that adds a new vector by merging a pair of budget vectors. The pair to merge is found by first searching for the budget vector with the smallest alpha-coefficients (measured in 2-norm), and then finding the second one by computing a certain degradation measure. This is therefore linear in the size of the budget.

The method is an implementation of the merge strategy given in wang, crammer, vucetic: "Breaking the Curse of Kernelization: Budgeted Stochastic Gradient Descent for Large-Scale SVM Training" and owes very much to the implementation in BudgetedSVM.

Author

Aydin Demircioglu

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This is an budget strategy that simply project one of the budget vectors onto the others. To save time, the smallest vector (measured in 2-norm of the alpha-coefficients) will be selected for projection.

Author

Aydin Demircioglu

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This is an budget strategy that simply removes one of the budget vectors. Depending on the flavor, this can be e.g. a random one, the smallest one (w.r.t. to 2-norm of the alphas)

Author

Aydin Demircioglu

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers, C. Igel

Date

2010, 2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

2013

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This implementation is based on a class removed from the LinAlg package, written by M. Kreutz in 1998.

Author

T. Glasmachers

Date

2007-2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause

Date

2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

B. Li

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

2010, 2013

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers,O.Krause

Date

2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

2010

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

M. Tuma

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

2011-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

K. N. Hansen, J. Kremer

Date

2011-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

2010-2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This file serves as a minimal abstraction layer. Inclusion of this file makes some frequently used functions, constants, and header file inclusions OS-, compiler-, and version-independent.

Author

-

Date

-

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Voss, M. Tuma

Date

2010

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

A ProxyReference can be used in the context of abstract functions to bind several related types of arguments to a single proxy type. Main use are ublas expression templates so that vectors, matrix rows and subvectors can be treated as one argument type

Author

O.Krause

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Implementation is based on

http://www.boost.org/doc/libs/1_52_0/doc/html/proto/appendices.html

Author

O. Krause

Date

2013

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Oswin Krause

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This class provides RAII handle management to Shark

Author

B. Li

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

The most important application of the methods provided in this file is the import of data from CSV files into Shark data containers.

Author

T. Voss, M. Tuma

Date

2010

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause

Date

2010-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

2006-2013

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This file provides containers for data used by the models, loss functions, and learning algorithms (trainers). The reason for dedicated containers of this type is that data often need to be split into subsets, such as training and test data, or folds in cross-validation. The containers in this file provide memory efficient mechanisms for managing and providing such subsets.

Author

O. Krause, T. Glasmachers

Date

2010-2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

The methods in this file allow to download data sets from the mldata.org repository and other sources.

Author

T. Glasmachers

Date

2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

The most important application of the methods provided in this file is the import of data from HDF5 files into Shark data containers.

Author

B. Li

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Aydin Demircioglu

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Deprecated:

This file is provided for backwards compatibility. Its is deprecated, use SparseData.h for new projects.

Author

T. Glasmachers

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

C. Igel

Date

2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

The most important application of the methods provided in this file is the import of data from LIBSVM files to Shark Data containers.

Author

M. Tuma, T. Glasmachers, C. Igel

Date

2010-2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause, C. Igel

Date

2010-2013

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This file provides containers for data used by the models, loss functions, and learning algorithms (trainers). The reason for dedicated containers of this type is that data often need to be split into subsets, such as training and test data, or folds in cross-validation. The containers in this file provide memory efficient mechanisms for managing and providing such subsets. The speciality of these containers are that they are weighted.

Author

O. Krause

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause, T.Glasmachers, T. Voss

Date

2010-2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/. Shark linear algebra definitions

This file provides all basic definitions for linear algebra. If defines objects and views for vectors and matrices over several base types as well as a lot of usefull functions.

Author

T. Glasmachers

Date

2007-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers

Date

2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers, A. Demircioglu

Date

2007-2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause

Date

2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Glasmachers, O. Krause

Date

2010

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause

Date

2010-01-01

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause

Date

2010-2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Date

2017

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Glasmachers, O. Krause, M. Tuma

Date

2010-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This class provides two main member functions for computing the derivative of a C-SVM hypothesis w.r.t. its hyperparameters. First, the derivative is prepared in general. Then, the derivative can be computed comparatively cheaply for any input sample. Needs to be supplied with pointers to a KernelExpansion and CSvmTrainer.

Author

M. Tuma, T. Glasmachers

Date

2007-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Affine linear kernel expansions resulting from Support vector machine (SVM) training and other kernel methods.

Author

T. Glasmachers

Date

2007-2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause

Date

2007-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Glasmachers, O. Krause, M. Tuma

Date

2010, 2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

M. Tuma, O.Krause

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Glasmachers, O. Krause, M. Tuma

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers, O.Krause

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Glasmachers

Date

2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

M. Tuma, T. Glasmachers, O. Krause

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

S., O.Krause

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers, O. Krause

Date

2010-2017

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Kang Li, O. Krause

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers, C. Igel, O.Krause

Date

2012-2017

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause

Date

2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

K. N. Hansen, J. Kremer

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

K. N. Hansen, O.Krause, J. Kremer

Date

2011-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause

Date

2013

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

-

Date

-

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

The function is described in

Christian Igel, Nikolaus Hansen, and Stefan Roth. Covariance Matrix Adaptation for Multi-objective Optimization. Evolutionary Computation 15(1), pp. 1-28, 2007

Author

-

Date

-

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

-

Date

2010-2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

-

Date

2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Multi-modal benchmark function.

Author

-

Date

-

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

The function is described in

H. Li and Q. Zhang. Multiobjective Optimization Problems with Complicated Pareto Sets, MOEA/D and NSGA-II, IEEE Trans on Evolutionary Computation, 2(12):284-302, April 2009.

Author

-

Date

-

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Class for balancing one or two poles on a cart using a fitness function that decreases the longer the pole(s) balance(s). Based on code written by Verena Heidrich-Meisner for the paper

V. Heidrich-Meisner and C. Igel. Neuroevolution strategies for episodic reinforcement learning. Journal of Algorithms, 64(4):152–168, 2009.

Author

Johan Valentin Damgaard

Date

-

Copyright 1995-2017 Shark Development Team

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Oswin Krause

Date

2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Class for simulating two poles balancing on a cart. Based on code written by Verena Heidrich-Meisner for the paper

V. Heidrich-Meisner and C. Igel. Neuroevolution strategies for episodic reinforcement learning. Journal of Algorithms, 64(4):152–168, 2009.

which was in turn based on code available at http://webdocs.cs.ualberta.ca/~sutton/book/code/pole.c as of 2015/4/19, written by Rich Sutton and Chuck Anderson and later modified. Faustino Gomez wrote the physics code using the differential equations from Alexis Weiland's paper and added the Runge-Kutta solver.

Author

Johan Valentin Damgaard

Date

-

Copyright 1995-2017 Shark Development Team

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Class for simulating a single pole balancing on a cart. Based on code written by Verena Heidrich-Meisner for the paper

V. Heidrich-Meisner and C. Igel. Neuroevolution strategies for episodic reinforcement learning. Journal of Algorithms, 64(4):152–168, 2009.

which was in turn based on code available at http://webdocs.cs.ualberta.ca/~sutton/book/code/pole.c as of 2015/4/19, written by Rich Sutton and Chuck Anderson and later modified. Faustino Gomez wrote the physics code using the differential equations from Alexis Weiland's paper and added the Runge-Kutta solver.

Author

Johan Valentin Damgaard

Date

-

Copyright 1995-2017 Shark Development Team

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

This non-convex benchmark function for real-valued optimization is a generalization from two to multiple dimensions of a classic function first proposed in:

H. H. Rosenbrock. An automatic method for finding the greatest or least value of a function. The Computer Journal 3: 175-184, 1960

Author

-

Date

-

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Voss

Date

2010-2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

The function is described in

Eckart Zitzler, Kalyanmoy Deb, and Lothar Thiele. Comparison of Multiobjective Evolutionary Algorithms: Empirical Results. Evolutionary Computation 8(2):173-195, 2000

Author

-

Date

-

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers, O. Krause

Date

2007-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T. Glasmachers, O.Krause

Date

2010-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Tobias Glasmachers

Date

2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Oswin Krause

Date

2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Oswin Krause, Christian Igel

Date

2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Christian Igel

Date

2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

C. Igel, T. Glasmachers, O. Krause

Date

2007-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Glasmachers, O.Krause

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

M.Tuma, T.Glasmachers

Date

2009-2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

T.Voss, O.Krause

Date

2016

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause

Date

2010-2011

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O.Krause

Date

2015

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause, A.Fischer

Date

2012-2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Oswin Krause Asja Fischer

Date

1.2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Asja Fischer

Date

1. 2014

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

Asja Fischer

Date

2013

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Author

O. Krause, A.Fischer, K.Bruegge

Date

2012

Copyright 1995-2017 Shark Development Team

---

This file is part of Shark. http://shark-ml.org/

Shark is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

Shark is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with Shark. If not, see http://www.gnu.org/licenses/.

Typedef Documentation

BinaryCD

typedef ContrastiveDivergence<BinaryGibbsOperator> shark::BinaryCD

Definition at line 53 of file BinaryRBM.h.

BinaryGibbsChain

typedef MarkovChain<BinaryGibbsOperator> shark::BinaryGibbsChain

Definition at line 49 of file BinaryRBM.h.

BinaryGibbsOperator

typedef GibbsOperator<BinaryRBM> shark::BinaryGibbsOperator

Definition at line 48 of file BinaryRBM.h.

BinaryParallelTempering

typedef SingleChainApproximator<BinaryPTChain> shark::BinaryParallelTempering

Definition at line 54 of file BinaryRBM.h.

BinaryPCD

typedef MultiChainApproximator<BinaryGibbsChain> shark::BinaryPCD

Definition at line 52 of file BinaryRBM.h.

BinaryPTChain

typedef TemperedMarkovChain<BinaryGibbsOperator> shark::BinaryPTChain

Definition at line 50 of file BinaryRBM.h.

BinaryRBM

typedef RBM<BinaryLayer,BinaryLayer, random::rng_type> shark::BinaryRBM

Definition at line 47 of file BinaryRBM.h.

BinarySpace

typedef TwoStateSpace<0,1> shark::BinarySpace

Definition at line 82 of file TwoStateSpace.h.

BipolarCD

typedef ContrastiveDivergence<BipolarGibbsOperator> shark::BipolarCD

Definition at line 53 of file BipolarRBM.h.

BipolarGibbsChain

typedef MarkovChain<BipolarGibbsOperator> shark::BipolarGibbsChain

Definition at line 49 of file BipolarRBM.h.

BipolarGibbsOperator

typedef GibbsOperator<BipolarRBM> shark::BipolarGibbsOperator

Definition at line 48 of file BipolarRBM.h.

BipolarParallelTempering

typedef SingleChainApproximator<BipolarPTChain> shark::BipolarParallelTempering

Definition at line 54 of file BipolarRBM.h.

BipolarPCD

typedef MultiChainApproximator<BipolarGibbsChain> shark::BipolarPCD

Definition at line 52 of file BipolarRBM.h.

BipolarPTChain

typedef TemperedMarkovChain<BipolarGibbsOperator> shark::BipolarPTChain

Definition at line 50 of file BipolarRBM.h.

BipolarRBM

typedef RBM<BipolarLayer,BipolarLayer, random::rng_type> shark::BipolarRBM

Definition at line 47 of file BipolarRBM.h.

ClassificationDataset

typedef LabeledData<RealVector, unsigned int> shark::ClassificationDataset

specialized template for classification with unsigned int labels

Definition at line 745 of file Dataset.h.

CompressedARDKernel

typedef ARDKernelUnconstrained<CompressedRealVector> shark::CompressedARDKernel

Definition at line 263 of file ArdKernel.h.

CompressedClassificationDataset

typedef LabeledData<CompressedRealVector, unsigned int> shark::CompressedClassificationDataset

specialized template for classification with unsigned int labels and sparse data

Definition at line 751 of file Dataset.h.

CompressedLinearKernel

typedef LinearKernel<CompressedRealVector> shark::CompressedLinearKernel

Definition at line 133 of file LinearKernel.h.

CompressedModelKernel

typedef ModelKernel<CompressedRealVector> shark::CompressedModelKernel

Definition at line 275 of file ModelKernel.h.

CompressedMonomialKernel

typedef MonomialKernel<CompressedRealVector> shark::CompressedMonomialKernel

Definition at line 193 of file MonomialKernel.h.

CompressedNormalizedKernel

typedef NormalizedKernel<CompressedRealVector> shark::CompressedNormalizedKernel

Definition at line 277 of file NormalizedKernel.h.

CompressedPolynomialKernel

typedef PolynomialKernel<CompressedRealVector> shark::CompressedPolynomialKernel

Definition at line 303 of file PolynomialKernel.h.

CompressedRbfKernel

typedef GaussianRbfKernel<CompressedRealVector> shark::CompressedRbfKernel

Definition at line 279 of file GaussianRbfKernel.h.

CompressedScaledKernel

typedef ScaledKernel<CompressedRealVector> shark::CompressedScaledKernel

Definition at line 160 of file ScaledKernel.h.

CompressedWeightedSumKernel

typedef WeightedSumKernel<CompressedRealVector> shark::CompressedWeightedSumKernel

Definition at line 399 of file WeightedSumKernel.h.

CompressesSubrangeKernel

typedef SubrangeKernel<CompressedRealVector> shark::CompressesSubrangeKernel

Definition at line 208 of file SubrangeKernel.h.

CrowdingRealCodedNSGAII

typedef IndicatorBasedRealCodedNSGAII< CrowdingDistance > shark::CrowdingRealCodedNSGAII

Definition at line 293 of file RealCodedNSGAII.h.

DenseARDKernel

typedef ARDKernelUnconstrained shark::DenseARDKernel

Definition at line 262 of file ArdKernel.h.

DenseLinearKernel

typedef LinearKernel shark::DenseLinearKernel

Definition at line 132 of file LinearKernel.h.

DenseModelKernel

typedef ModelKernel<RealVector> shark::DenseModelKernel

Definition at line 274 of file ModelKernel.h.

DenseMonomialKernel

typedef MonomialKernel shark::DenseMonomialKernel

Definition at line 192 of file MonomialKernel.h.

DenseNormalizedKernel

typedef NormalizedKernel shark::DenseNormalizedKernel

Definition at line 276 of file NormalizedKernel.h.

DensePolynomialKernel

typedef PolynomialKernel shark::DensePolynomialKernel

Definition at line 302 of file PolynomialKernel.h.

DenseRbfKernel

typedef GaussianRbfKernel shark::DenseRbfKernel

Definition at line 278 of file GaussianRbfKernel.h.

DenseScaledKernel

typedef ScaledKernel shark::DenseScaledKernel

Definition at line 159 of file ScaledKernel.h.

DenseSubrangeKernel

typedef SubrangeKernel<RealVector> shark::DenseSubrangeKernel

Definition at line 207 of file SubrangeKernel.h.

DenseWeightedSumKernel

typedef WeightedSumKernel shark::DenseWeightedSumKernel

Definition at line 398 of file WeightedSumKernel.h.

EpsilonMOCMA

typedef IndicatorBasedMOCMA< AdditiveEpsilonIndicator > shark::EpsilonMOCMA

Definition at line 265 of file MOCMA.h.

EpsilonSteadyStateMOCMA

typedef IndicatorBasedSteadyStateMOCMA< AdditiveEpsilonIndicator > shark::EpsilonSteadyStateMOCMA

Definition at line 286 of file SteadyStateMOCMA.h.

EpsRealCodedNSGAII

typedef IndicatorBasedRealCodedNSGAII< AdditiveEpsilonIndicator > shark::EpsRealCodedNSGAII

Definition at line 292 of file RealCodedNSGAII.h.

FrontType

typedef std::vector< shark::RealVector > shark::FrontType

Definition at line 14 of file AdditiveEpsilonIndicatorMain.cpp.

GaussianBinaryCD

typedef ContrastiveDivergence<GaussianBinaryGibbsOperator> shark::GaussianBinaryCD

Definition at line 54 of file GaussianBinaryRBM.h.

GaussianBinaryGibbsChain

typedef MarkovChain<GaussianBinaryGibbsOperator> shark::GaussianBinaryGibbsChain

Definition at line 50 of file GaussianBinaryRBM.h.

GaussianBinaryGibbsOperator

typedef GibbsOperator<GaussianBinaryRBM> shark::GaussianBinaryGibbsOperator

Definition at line 49 of file GaussianBinaryRBM.h.

GaussianBinaryParallelTempering

typedef SingleChainApproximator<GaussianBinaryPTChain> shark::GaussianBinaryParallelTempering

Definition at line 55 of file GaussianBinaryRBM.h.

GaussianBinaryPCD

typedef MultiChainApproximator<GaussianBinaryGibbsChain> shark::GaussianBinaryPCD

Definition at line 53 of file GaussianBinaryRBM.h.

GaussianBinaryPTChain

typedef TemperedMarkovChain<GaussianBinaryGibbsOperator> shark::GaussianBinaryPTChain

Definition at line 51 of file GaussianBinaryRBM.h.

GaussianBinaryRBM

typedef RBM<GaussianLayer,BinaryLayer, random::rng_type> shark::GaussianBinaryRBM

Definition at line 48 of file GaussianBinaryRBM.h.

InArchive

typedef boost::archive::polymorphic_iarchive shark::InArchive

Type of an archive to read from.

Definition at line 74 of file ISerializable.h.

MOCMA

typedef IndicatorBasedMOCMA< HypervolumeIndicator > shark::MOCMA

Definition at line 264 of file MOCMA.h.

MultiObjectiveFunction

typedef AbstractObjectiveFunction< RealVector, RealVector > shark::MultiObjectiveFunction

Definition at line 318 of file AbstractObjectiveFunction.h.

OutArchive

typedef boost::archive::polymorphic_oarchive shark::OutArchive

Type of an archive to write to.

Definition at line 80 of file ISerializable.h.

PermutationMatrix

typedef blas::permutation_matrix shark::PermutationMatrix

Definition at line 74 of file Base.h.

Preference

typedef std::pair<double, RealVector> shark::Preference

A preferred region in a lattice-sampled unit sphere.

A preferred region is a pair with a radius and a vector. The intersection of the vector and the unit sphere denotes the center of the preferred region, and the radius denotes the radius of a sphere constructed such that the center point (intersection of vector and unit sphere) is on its periphery. Points are then sampled on this smaller sphere and projected up on the unit sphere, thus restricting the covered segment/area/volume/hypervolume of the unit sphere. See Figure 1 in "Evolutionary Many-objective Optimization of Hybrid Electric Vehicle Control: From General Optimization to Preference Articulation"

Definition at line 136 of file Lattice.h.

RealCodedNSGAII

typedef IndicatorBasedRealCodedNSGAII< HypervolumeIndicator > shark::RealCodedNSGAII

Definition at line 291 of file RealCodedNSGAII.h.

RecreationIndices

typedef std::pair< std::vector<std::size_t> , std::vector<std::size_t> > shark::RecreationIndices

auxiliary typedef for createCVSameSizeBalanced and createCVFullyIndexed, stores location index in the first and partition index in the second

Definition at line 127 of file CVDatasetTools.h.

RegressionDataset

typedef LabeledData<RealVector, RealVector> shark::RegressionDataset

specialized template for regression with RealVector labels

Definition at line 748 of file Dataset.h.

Rprop93

typedef RpropPlus shark::Rprop93

Used to connect the class names with the year of publication of the paper in which the algorithm was introduced.

Definition at line 539 of file Rprop.h.

Rprop94

typedef RpropMinus shark::Rprop94

Used to connect the class names with the year of publication of the paper in which the algorithm was introduced.

Definition at line 543 of file Rprop.h.

Rprop99

typedef IRpropPlus shark::Rprop99

Used to connect the class names with the year of publication of the paper in which the algorithm was introduced.

Definition at line 531 of file Rprop.h.

Rprop99d

typedef IRpropMinus shark::Rprop99d

Used to connect the class names with the year of publication of the paper in which the algorithm was introduced.

Definition at line 535 of file Rprop.h.

Sequence

typedef std::deque<RealVector> shark::Sequence

Type of Data sequences.

Definition at line 77 of file Base.h.

SingleObjectiveFunction

typedef AbstractObjectiveFunction< RealVector, double > shark::SingleObjectiveFunction

Definition at line 317 of file AbstractObjectiveFunction.h.

SteadyStateMOCMA

typedef IndicatorBasedSteadyStateMOCMA< HypervolumeIndicator > shark::SteadyStateMOCMA

Definition at line 285 of file SteadyStateMOCMA.h.

SymmetricBinarySpace

typedef TwoStateSpace<-1,1> shark::SymmetricBinarySpace

Definition at line 83 of file TwoStateSpace.h.

TextInArchive

typedef boost::archive::polymorphic_text_iarchive shark::TextInArchive

Definition at line 75 of file ISerializable.h.

TextOutArchive

typedef boost::archive::polymorphic_text_oarchive shark::TextOutArchive

Definition at line 81 of file ISerializable.h.

TruncExpBinaryCD

typedef ContrastiveDivergence<TruncExpBinaryGibbsOperator> shark::TruncExpBinaryCD

Definition at line 53 of file TruncExpBinaryRBM.h.

TruncExpBinaryGibbsChain

typedef MarkovChain<TruncExpBinaryGibbsOperator> shark::TruncExpBinaryGibbsChain

Definition at line 49 of file TruncExpBinaryRBM.h.

TruncExpBinaryGibbsOperator

typedef GibbsOperator<TruncExpBinaryRBM> shark::TruncExpBinaryGibbsOperator

Definition at line 48 of file TruncExpBinaryRBM.h.

TruncExpBinaryParallelTempering

typedef SingleChainApproximator<TruncExpBinaryPTChain> shark::TruncExpBinaryParallelTempering

Definition at line 54 of file TruncExpBinaryRBM.h.

TruncExpBinaryPCD

typedef MultiChainApproximator<TruncExpBinaryGibbsChain> shark::TruncExpBinaryPCD

Definition at line 52 of file TruncExpBinaryRBM.h.

TruncExpBinaryPTChain

typedef TemperedMarkovChain<TruncExpBinaryGibbsOperator> shark::TruncExpBinaryPTChain

Definition at line 50 of file TruncExpBinaryRBM.h.

TruncExpBinaryRBM

typedef RBM<TruncExpBinaryEnergy, random::rng_type> shark::TruncExpBinaryRBM

Definition at line 47 of file TruncExpBinaryRBM.h.

Enumeration Type Documentation

AlphaStatus

enum shark::AlphaStatus

Enumerator
AlphaFree
AlphaLowerBound
AlphaUpperBound
AlphaDeactivated

Definition at line 391 of file QpSolver.h.

BuildType

enum shark::BuildType

Models the build type.

Enumerator
RELEASE_BUILD_TYPE	A release build.
DEBUG_BUILD_TYPE	A debug build.

Definition at line 58 of file Shark.h.

Convolution

enum shark::Convolution

strong

Enumerator
Valid
ZeroPad

Definition at line 42 of file ConvolutionalModel.h.

DominanceRelation

enum shark::DominanceRelation

Result of comparing two objective vectors w.r.t. Pareto dominance.

Enumerator
INCOMPARABLE
LHS_DOMINATES_RHS
RHS_DOMINATES_LHS
EQUIVALENT

Definition at line 42 of file ParetoDominance.h.

McSvm

enum shark::McSvm

strong

Enumerator
WW
CS
LLW
ATM
ATS
ADM
OVA
MMR
ReinforcedSvm

Definition at line 27 of file CSvmTrainer.h.

PartitionEstimationAlgorithm

enum shark::PartitionEstimationAlgorithm

Enumerator
AIS
AISMean
TwoSidedAISMean
AcceptanceRatio
AcceptanceRatioMean

Definition at line 111 of file analytics.h.

QpStopType

enum shark::QpStopType

reason for the quadratic programming solver to stop the iterative optimization process

Enumerator
QpNone
QpAccuracyReached
QpMaxIterationsReached
QpTimeout

Definition at line 63 of file QuadraticProgram.h.

Function Documentation

annealedImportanceSampling() [1/2]

template<class RBMType >

double shark::annealedImportanceSampling	(	RBMType &	rbm,
		RealVector const &	beta,
		std::size_t	samples
	)

Definition at line 171 of file analytics.h.

Referenced by annealedImportanceSampling().

annealedImportanceSampling() [2/2]

template<class RBMType >

double shark::annealedImportanceSampling	(	RBMType &	rbm,
		std::size_t	chains,
		std::size_t	samples
	)

Definition at line 196 of file analytics.h.

References annealedImportanceSampling(), and beta.

approximateKernelExpansion()

SHARK_EXPORT_SYMBOL KernelExpansion<RealVector> shark::approximateKernelExpansion	(	random::rng_type &	rng,
		KernelExpansion< RealVector > const &	model,
		std::size_t	k,
		double	precision = 1.e-8
	)

Approximates a kernel expansion by a smaller one using an optimized basis.

often, a kernel expansion can be represented much more compactly when the points defining the basis of the kernel expansion are not fixed. The resulting kernel expansion can be evaluated much quicker than the original when k is small compared to the number of the nonzero elements in the weight vector of the supplied kernel expansion.

Given a kernel expansion with weight matrix alpha and Basis B of size m, finds a new weight matrix beta and Basis Z with k vectors so that the difference of the resulting decision vectors is small in the RKHS defined by the supplied kernel.

The algorithm proceeds by first performing a kMeans clustering as a good initialization. This initial guess is then optimized by finding the closest weight vector to the original vector representable by the basis. Using this estimate, the basis can then be optimized.

The supplied kernel must be dereferentiable wrt its input parameters which excludes all kernels not defined on RealVector

The algorithms is O(k^3 + k m) in each iteration.

Parameters

rng	the Rng used for the kMeans clustering
model	the kernel expansion to approximate
k	the number of basis vectors to be used by the approximation
precision	target precision of the gradient to be reached during optimization

batchBegin() [1/2]

template<class BatchT >

auto shark::batchBegin

(

BatchT &

batch

)

-> decltype(BatchTraits<BatchT>::type::begin(batch))

Definition at line 420 of file BatchInterface.h.

Referenced by transform().

batchBegin() [2/2]

template<class BatchT >

auto shark::batchBegin

(

BatchT const &

batch

)

-> decltype(BatchTraits<BatchT>::type::begin(batch))

Definition at line 425 of file BatchInterface.h.

batchEnd() [1/2]

template<class BatchT >

auto shark::batchEnd

(

BatchT &

batch

)

-> decltype(BatchTraits<BatchT>::type::end(batch))

Definition at line 430 of file BatchInterface.h.

Referenced by shark::ContrastiveDivergence< Operator >::evalDerivative(), and transform().

batchEnd() [2/2]

template<class BatchT >

auto shark::batchEnd

(

BatchT const &

batch

)

-> decltype(BatchTraits<BatchT>::type::end(batch))

Definition at line 435 of file BatchInterface.h.

batchSize()

template<class BatchT >

std::size_t shark::batchSize

(

BatchT const &

batch

)

Definition at line 415 of file BatchInterface.h.

Referenced by calculateKernelMatrixParameterDerivative(), calculateMixedKernelMatrix(), shark::KHCTree< Container, CuttingAccuracy >::calculateNormal(), calculateRegularizedKernelMatrix(), createCVFullyIndexed(), createCVIID(), createCVIndexed(), createCVSameSize(), createCVSameSizeBalanced(), createLabeledDataFromRange(), shark::GibbsOperator< RBMType >::createSample(), createUnlabeledDataFromRange(), shark::DataView< shark::Data< LabelType > const >::DataView(), downloadFromMLData(), downloadSparseData(), shark::Energy< RBM >::energy(), shark::Energy< RBM >::energyFromHiddenInput(), shark::Energy< RBM >::energyFromVisibleInput(), shark::GaussianLayer::energyTerm(), shark::MissingFeaturesKernelExpansion< InputType >::eval(), shark::PointSetKernel< InputType >::eval(), shark::Classifier< KernelExpansion< InputType > >::eval(), shark::DiscreteKernel::eval(), shark::NormalizedKernel< InputType >::eval(), shark::ProductKernel< MultiTaskSample< InputTypeT > >::eval(), shark::OneVersusOneClassifier< InputType, VectorType >::eval(), shark::WeightedSumKernel< InputType >::eval(), shark::KernelExpansion< InputType >::eval(), shark::KernelTargetAlignment< InputType, LabelType >::evalDerivative(), shark::TruncatedExponentialLayer::expectedPhiValue(), exportSparseData(), shark::AbstractKernelFunction< MultiTaskSample< InputTypeT > >::featureDistanceSqr(), shark::ExactGradient< RBMType >::getLogPartition(), shark::SimpleNearestNeighbors< InputType, LabelType >::getNeighbors(), shark::TreeNearestNeighbors< InputType, LabelType >::getNeighbors(), shark::HierarchicalClustering< InputT >::hardMembership(), shark::AbstractClustering< RealVector >::hardMembership(), import_libsvm(), shark::Energy< RBM >::logUnnormalizedProbabilityHidden(), shark::Energy< RBM >::logUnnormalizedProbabilityVisible(), main(), repartitionByClass(), shark::MultiChainApproximator< MarkovChainType >::setBatchSize(), splitAtElement(), toDataset(), shark::ScaledKernel< InputType >::weightedInputDerivative(), shark::NormalizedKernel< InputType >::weightedInputDerivative(), shark::WeightedSumKernel< InputType >::weightedInputDerivative(), shark::PointSetKernel< InputType >::weightedParameterDerivative(), and shark::NormalizedKernel< InputType >::weightedParameterDerivative().

bootstrap() [1/2]

template<class InputType , class LabelType >

WeightedLabeledData< InputType, LabelType> shark::bootstrap	(	LabeledData< InputType, LabelType > const &	dataset,
		std::size_t	bootStrapSize = 0
	)

Creates a bootstrap partition of a labeled dataset and returns it using weighting.

Bootstrapping resamples the dataset by drawing a set of points with replacement. Thus the sampled set will contain some points multiple times and some points not at all. Bootstrapping is usefull to obtain unbiased measurements of the mean and variance of an estimator.

Optionally the size of the bootstrap (that is, the number of sampled points) can be set. By default it is 0, which indicates that it is the same size as the original dataset.

Definition at line 801 of file WeightedDataset.h.

References shark::random::discrete(), shark::random::globalRng, shark::LabeledData< InputT, LabelT >::inputShape(), shark::WeightedLabeledData< InputT, LabelT >::inputShape(), shark::LabeledData< InputT, LabelT >::labelShape(), shark::WeightedLabeledData< InputT, LabelT >::labelShape(), and shark::LabeledData< InputT, LabelT >::numberOfElements().

bootstrap() [2/2]

template<class InputType >

WeightedUnlabeledData<InputType> shark::bootstrap	(	UnlabeledData< InputType > const &	dataset,
		std::size_t	bootStrapSize = 0
	)

Creates a bootstrap partition of an unlabeled dataset and returns it using weighting.

Bootstrapping resamples the dataset by drawing a set of points with replacement. Thus the sampled set will contain some points multiple times and some points not at all. Bootstrapping is usefull to obtain unbiased measurements of the mean and variance of an estimator.

Optionally the size of the bootstrap (that is, the number of sampled points) can be set. By default it is 0, which indicates that it is the same size as the original dataset.

Definition at line 829 of file WeightedDataset.h.

References shark::random::discrete(), shark::random::globalRng, shark::Data< Type >::numberOfElements(), shark::Data< Type >::shape(), and shark::WeightedUnlabeledData< DataT >::shape().

calculateKernelMatrixParameterDerivative()

template<class InputType , class WeightMatrix >

RealVector shark::calculateKernelMatrixParameterDerivative	(	AbstractKernelFunction< InputType > const &	kernel,
		Data< InputType > const &	dataset,
		WeightMatrix const &	weights
	)

Efficiently calculates the weighted derivative of a Kernel Gram Matrix w.r.t the Kernel Parameters.

The formula is \( \sum_i \sum_j w_{ij} k(x_i,x_j)\) where w_ij are the weights of the gradient and x_i x_j are the datapoints defining the gram matrix and k is the kernel. For efficiency it is assumd that w_ij = w_ji. This method is only useful when the whole Kernel Gram Matrix neds to be computed to get the weights w_ij and only computing smaller blocks is not sufficient.

Parameters

kernel	the kernel for which to calculate the kernel gram matrix
dataset	the set of points used in the gram matrix
weights	the weights of the derivative, they must be symmetric!

Returns

the weighted derivative w.r.t the parameters.

Definition at line 181 of file KernelHelpers.h.

References shark::Data< Type >::batch(), batchSize(), shark::AbstractKernelFunction< InputTypeT >::createState(), shark::AbstractKernelFunction< InputTypeT >::eval(), shark::Data< Type >::numberOfBatches(), shark::IParameterizable< VectorType >::numberOfParameters(), and shark::AbstractKernelFunction< InputTypeT >::weightedParameterDerivative().

Referenced by shark::RadiusMarginQuotient< InputType, CacheType >::evalDerivative().

calculateMixedKernelMatrix() [1/2]

template<class InputType , class M , class Device >

void shark::calculateMixedKernelMatrix	(	AbstractKernelFunction< InputType >const &	kernel,
		Data< InputType > const &	dataset1,
		Data< InputType > const &	dataset2,
		blas::matrix_expression< M, Device > &	matrix
	)

Calculates the kernel gram matrix between two data sets.

Parameters

kernel	the kernel for which to calculate the kernel gram matrix
dataset1	the set of points corresponding to rows of the Gram matrix
dataset2	the set of points corresponding to columns of the Gram matrix
matrix	the target kernel matrix

Definition at line 95 of file KernelHelpers.h.

References shark::Data< Type >::batch(), batchSize(), shark::Data< Type >::numberOfBatches(), shark::Data< Type >::numberOfElements(), SHARK_PARALLEL_FOR, and SIZE_CHECK.

Referenced by calculateMixedKernelMatrix().

calculateMixedKernelMatrix() [2/2]

template<class InputType >

RealMatrix shark::calculateMixedKernelMatrix	(	AbstractKernelFunction< InputType >const &	kernel,
		Data< InputType > const &	dataset1,
		Data< InputType > const &	dataset2
	)

Calculates the kernel gram matrix between two data sets.

Parameters

kernel	the kernel for which to calculate the kernel gram matrix
dataset1	the set of points corresponding to rows of the Gram matrix
dataset2	the set of points corresponding to columns of the Gram matrix

Returns

matrix the target kernel matrix

Definition at line 159 of file KernelHelpers.h.

References calculateMixedKernelMatrix().

calculateRegularizedKernelMatrix() [1/2]

template<class InputType , class M , class Device >

void shark::calculateRegularizedKernelMatrix	(	AbstractKernelFunction< InputType >const &	kernel,
		Data< InputType > const &	dataset,
		blas::matrix_expression< M, Device > &	matrix,
		double	regularizer = 0
	)

Calculates the regularized kernel gram matrix of the points stored inside a dataset.

Regularization is applied by adding the regularizer on the diagonal

Parameters

kernel	the kernel for which to calculate the kernel gram matrix
dataset	the set of points used in the gram matrix
matrix	the target kernel matrix
regularizer	the regularizer of the matrix which is always >= 0. default is 0.

Definition at line 52 of file KernelHelpers.h.

References shark::Data< Type >::batch(), batchSize(), shark::Data< Type >::numberOfBatches(), shark::Data< Type >::numberOfElements(), SHARK_PARALLEL_FOR, SHARK_RUNTIME_CHECK, and SIZE_CHECK.

Referenced by calculateRegularizedKernelMatrix(), shark::KernelMatrix< InputType, CacheType >::matrix(), and shark::RegularizationNetworkTrainer< InputType >::train().

calculateRegularizedKernelMatrix() [2/2]

template<class InputType >

RealMatrix shark::calculateRegularizedKernelMatrix	(	AbstractKernelFunction< InputType >const &	kernel,
		Data< InputType > const &	dataset,
		double	regularizer = 0
	)

Calculates the regularized kernel gram matrix of the points stored inside a dataset.

Regularization is applied by adding the regularizer on the diagonal

Parameters

kernel	the kernel for which to calculate the kernel gram matrix
dataset	the set of points used in the gram matrix
regularizer	the regularizer of the matrix which is always >= 0. default is 0.

Returns

the kernel gram matrix

Definition at line 141 of file KernelHelpers.h.

References calculateRegularizedKernelMatrix(), and SHARK_RUNTIME_CHECK.

classSizes() [1/2]

std::vector<std::size_t> shark::classSizes ( WeightedUnlabeledData< unsigned int > const &  labels )

inline

Returns the number of members of each class in the dataset.

Definition at line 701 of file WeightedDataset.h.

References classSizes().

classSizes() [2/2]

template<class InputType , class LabelType >

std::vector<std::size_t> shark::classSizes ( WeightedLabeledData< InputType, LabelType > const &  dataset )

inline

Returns the number of members of each class in the dataset.

Definition at line 730 of file WeightedDataset.h.

References classSizes(), and shark::WeightedLabeledData< InputT, LabelT >::labels().

classWeight()

template<class InputType >

RealVector shark::classWeight

(

WeightedLabeledData< InputType, unsigned int > const &

dataset

)

Computes the cumulative weight of every class.

Definition at line 755 of file WeightedDataset.h.

References numberOfClasses().

Referenced by shark::KernelMeanClassifier< InputType >::train().

computeClosestNeighbourIndicesOnLattice()

template<typename Matrix >

UIntMatrix shark::computeClosestNeighbourIndicesOnLattice	(	Matrix const &	m,
		std::size_t const	n
	)

Definition at line 175 of file Lattice.h.

computeOptimalLatticeTicks()

std::size_t shark::computeOptimalLatticeTicks	(	std::size_t const	n,
		std::size_t const	target_count
	)

Computes the number of Ticks for a grid of a certain size.

Computes the least number of ticks in each dimension required to make an n-dimensional simplex grid at least as many points as a target number points. For example, the points in a two-dimensional grid – a line – with size n are the points (0,n-1), (1,n-2), ... (n-1,0).

Referenced by shark::NSGA3Indicator::init().

covariance()

template<class VectorType >

blas::matrix<typename VectorType::value_type> shark::covariance

(

Data< VectorType > const &

data

)

Calculates the covariance matrix of the data vectors.

Referenced by createData(), mean(), shark::NormalDistributedPoints::NormalDistributedPoints(), and shark::NormalizeComponentsZCA::train().

createBatch() [1/3]

template<class T , class Range >

Batch<T>::type shark::createBatch

(

Range const &

range

)

creates a batch from a range of inputs

Definition at line 389 of file BatchInterface.h.

Referenced by shark::AbstractLoss< unsigned int, RealVector >::eval(), shark::AbstractKernelFunction< MultiTaskSample< InputTypeT > >::eval(), shark::NormalizedKernel< InputType >::eval(), shark::AbstractModel< InputT, unsigned int >::eval(), shark::AbstractLoss< unsigned int, RealVector >::evalDerivative(), shark::AbstractClustering< RealVector >::hardMembership(), shark::CSvmDerivative< InputType, CacheType >::modelCSvmParameterDerivative(), subBatch(), shark::KernelSGDTrainer< InputType, CacheType >::train(), shark::KernelBudgetedSGDTrainer< InputType, CacheType >::train(), and shark::CSvmDerivative< InputType, CacheType >::write().

createBatch() [2/3]

template<class Range >

Batch<typename Range::value_type>::type shark::createBatch

(

Range const &

range

)

creates a batch from a range of inputs

Definition at line 395 of file BatchInterface.h.

createBatch() [3/3]

template<class T , class Iterator >

Batch<T>::type shark::createBatch	(	Iterator const &	begin,
		Iterator const &	end
	)

Definition at line 400 of file BatchInterface.h.

createLabeledDataFromRange()

template<class InputRange , class LabelRange , class WeightRange >

boost::disable_if< boost::is_arithmetic<WeightRange>, WeightedLabeledData< typename boost::range_value<InputRange>::type, typename boost::range_value<LabelRange>::type >>::type shark::createLabeledDataFromRange	(	InputRange const &	inputs,
		LabelRange const &	labels,
		WeightRange const &	weights,
		std::size_t	batchSize = 0
	)

creates a weighted unweighted data object from two ranges, representing data and weights

Definition at line 773 of file WeightedDataset.h.

References batchSize(), createDataFromRange(), createLabeledDataFromRange(), and SHARK_RUNTIME_CHECK.

createUnlabeledDataFromRange()

template<class DataRange , class WeightRange >

boost::disable_if< boost::is_arithmetic<WeightRange>, WeightedUnlabeledData< typename boost::range_value<DataRange>::type > >::type shark::createUnlabeledDataFromRange	(	DataRange const &	data,
		WeightRange const &	weights,
		std::size_t	batchSize = 0
	)

creates a weighted unweighted data object from two ranges, representing data and weights

Definition at line 544 of file WeightedDataset.h.

References batchSize(), createDataFromRange(), createUnlabeledDataFromRange(), and SHARK_RUNTIME_CHECK.

dataDimension() [1/2]

template<class DatasetType >

std::size_t shark::dataDimension

(

DataView< DatasetType > const &

view

)

Return the dimensionality of the dataset represented by the view.

Definition at line 344 of file DataView.h.

References dataDimension(), and shark::DataView< DatasetType >::dataset().

dataDimension() [2/2]

template<class InputType >

std::size_t shark::dataDimension

(

WeightedUnlabeledData< InputType > const &

dataset

)

Return the dimnsionality of points of a weighted dataset.

Definition at line 707 of file WeightedDataset.h.

References dataDimension().

dcNonDominatedSort()

template<class PointRange , class RankRange >

void shark::dcNonDominatedSort	(	PointRange const &	points,
		RankRange &	ranks
	)

Definition at line 466 of file DCNonDominatedSort.h.

Referenced by nonDominatedSort().

dominance()

template<class VectorTypeA , class VectorTypeB >

DominanceRelation shark::dominance ( VectorTypeA const &  lhs, VectorTypeB const &  rhs  )

inline

Pareto dominance relation for two objective vectors.

Definition at line 52 of file ParetoDominance.h.

References EQUIVALENT, INCOMPARABLE, LHS_DOMINATES_RHS, RHS_DOMINATES_LHS, and SHARK_ASSERT.

Referenced by fastNonDominatedSort(), and shark::BaseDCNonDominatedSort::operator()().

estimateLogFreeEnergy() [1/2]

template<class RBMType >

double shark::estimateLogFreeEnergy	(	RBMType &	rbm,
		Data< RealVector > const &	initDataset,
		RealVector const &	beta,
		std::size_t	samples,
		PartitionEstimationAlgorithm	algorithm = AcceptanceRatioMean,
		float	burnInPercentage = 0.1
	)

Definition at line 154 of file analytics.h.

References estimateLogFreeEnergyFromEnergySamples().

Referenced by estimateLogFreeEnergy().

estimateLogFreeEnergy() [2/2]

template<class RBMType >

double shark::estimateLogFreeEnergy	(	RBMType &	rbm,
		Data< RealVector > const &	initDataset,
		std::size_t	chains,
		std::size_t	samples,
		PartitionEstimationAlgorithm	algorithm = AIS,
		float	burnInPercentage = 0.1
	)

Definition at line 182 of file analytics.h.

References beta, and estimateLogFreeEnergy().

estimateLogFreeEnergyFromEnergySamples()

double shark::estimateLogFreeEnergyFromEnergySamples ( RealMatrix const &  energyDiffUp, RealMatrix const &  energyDiffDown, PartitionEstimationAlgorithm  algorithm = AIS  )

inline

Definition at line 119 of file analytics.h.

References AIS.

Referenced by estimateLogFreeEnergy().

evalSkipMissingFeatures() [1/2]

template<typename InputType , typename InputTypeT1 , typename InputTypeT2 >

double shark::evalSkipMissingFeatures	(	const AbstractKernelFunction< InputType > &	kernelFunction,
		const InputTypeT1 &	inputA,
		const InputTypeT2 &	inputB
	)

Does a kernel function evaluation with Missing features in the inputs

Parameters

kernelFunction	The kernel function used to do evaluation
inputA	a input
inputB	another input

The kernel k(x,y) is evaluated taking missing features into account. For this it is checked whether a feature of x or y is nan and in this case the corresponding features in inputA and inputB won't be considered.

Definition at line 58 of file EvalSkipMissingFeatures.h.

References shark::AbstractKernelFunction< InputTypeT >::eval(), SHARK_RUNTIME_CHECK, SIZE_CHECK, and shark::AbstractKernelFunction< InputTypeT >::supportsVariableInputSize().

Referenced by shark::ExampleModifiedKernelMatrix< InputType, CacheType >::entry().

evalSkipMissingFeatures() [2/2]

template<typename InputType , typename InputTypeT1 , typename InputTypeT2 , typename InputTypeT3 >

double shark::evalSkipMissingFeatures	(	const AbstractKernelFunction< InputType > &	kernelFunction,
		const InputTypeT1 &	inputA,
		const InputTypeT2 &	inputB,
		InputTypeT3 const &	missingness
	)

Do kernel function evaluation while Missing features in the inputs

Parameters

kernelFunction	The kernel function used to do evaluation
inputA	a input
inputB	another input
missingness	used to decide which features in the inputs to take into consideration for the purpose of evaluation. If a feature is NaN, then the corresponding features in inputA and inputB won't be considered.

Definition at line 104 of file EvalSkipMissingFeatures.h.

References shark::AbstractKernelFunction< InputTypeT >::eval(), SHARK_RUNTIME_CHECK, SIZE_CHECK, and shark::AbstractKernelFunction< InputTypeT >::supportsVariableInputSize().

exportFiltersToPGMGrid() [1/2]

void shark::exportFiltersToPGMGrid ( std::string const &  basename, RealMatrix const &  filters, std::size_t  width, std::size_t  height  )

inline

Exports a set of filters as a grid image.

It is assumed that the filters each form a row in the filter-matrix. Moreover, the sizes of the filter images has to be given and it must gold width*height=W.size2(). The filters a re printed on a single image as a grid. The grid will be close to square. And the image are separated by a black 1 pixel wide line. The output will be normalized so that all images are on the same scale.

Parameters

basename	File to write to. ".pgm" is appended to the filename
filters	Matrix storing the filters row by row
width	Width of the filter image
height	Height of th filter image

Definition at line 146 of file Pgm.h.

References SIZE_CHECK.

Referenced by main().

exportFiltersToPGMGrid() [2/2]

void shark::exportFiltersToPGMGrid ( std::string const &  basename, Data< RealVector > const &  filters, std::size_t  width, std::size_t  height  )

inline

Exports a set of filters as a grid image.

It is assumed that the filters each form a row in the filter-matrix. Moreover, the sizes of the filter images has to be given and it must gold width*height=W.size2(). The filters a re printed on a single image as a grid. The grid will be close to square. And the image are separated by a black 1 pixel wide line. The output will be normalized so that all images are on the same scale.

Parameters

basename	File to write to. ".pgm" is appended to the filename
filters	Matrix storing the filters row by row
width	Width of the filter image
height	Height of th filter image

Definition at line 183 of file Pgm.h.

References dataDimension(), shark::Data< Type >::numberOfElements(), and SIZE_CHECK.

exportPGM()

template<class T >

void shark::exportPGM	(	std::string const &	fileName,
		T const &	data,
		std::size_t	sx,
		std::size_t	sy,
		bool	normalize = false
	)

Export a PGM image to file.

Parameters

fileName	File to write to
data	Linear object storing image
sx	Width of image
sy	Height of image
normalize	Adjust values to [0,255], default false

Definition at line 118 of file Pgm.h.

References SIZE_CHECK.

Referenced by main().

fastNonDominatedSort()

template<class PointRange , class RankRange >

void shark::fastNonDominatedSort	(	PointRange const &	points,
		RankRange &	ranks
	)

Implements the well-known non-dominated sorting algorithm.

Assembles subsets/fronts of mututally non-dominating individuals. Afterwards every individual is assigned a rank by points[i].rank() = frontNumber. The front of dominating points has the value 1.

The algorithm is dscribed in Deb et al, A Fast and Elitist Multiobjective Genetic Algorithm: NSGA-II IEEE Transactions on Evolutionary Computation, 2002

Parameters

points	[in] pointsulation to subdivide into fronts of non-dominated individuals.
ranks	[out] Set of integers storing the rank of the i-th point. must have the same size

Definition at line 49 of file FastNonDominatedSort.h.

References dominance(), LHS_DOMINATES_RHS, RHS_DOMINATES_LHS, and SIZE_CHECK.

Referenced by nonDominatedSort().

fusionize() [1/4]

template<class S >

S& shark::fusionize

(

detail::FusionFacade< S > &

facade

)

Definition at line 147 of file BatchInterfaceAdaptStruct.h.

References S.

fusionize() [2/4]

template<class S >

S const& shark::fusionize

(

detail::FusionFacade< S > const &

facade

)

Definition at line 151 of file BatchInterfaceAdaptStruct.h.

References S.

fusionize() [3/4]

template<class S >

boost::disable_if<detail::isFusionFacade<S>,S&>::type shark::fusionize

(

S &

facade

)

Definition at line 157 of file BatchInterfaceAdaptStruct.h.

References S.

fusionize() [4/4]

template<class S >

boost::disable_if<detail::isFusionFacade<S>,S const& >::type shark::fusionize

(

S const &

facade

)

Definition at line 162 of file BatchInterfaceAdaptStruct.h.

getBatchElement() [1/2]

template<class BatchT >

auto shark::getBatchElement	(	BatchT &	batch,
		std::size_t	i
	)		-> decltype(BatchTraits<BatchT>::type::get(std::declval<BatchT&>(),i))

Definition at line 405 of file BatchInterface.h.

Referenced by shark::DifferenceKernelMatrix< InputType, CacheType >::entry(), shark::AbstractLoss< unsigned int, RealVector >::eval(), shark::PointSetKernel< InputType >::eval(), shark::AbstractKernelFunction< MultiTaskSample< InputTypeT > >::eval(), shark::NormalizedKernel< InputType >::eval(), shark::AbstractModel< InputT, unsigned int >::eval(), shark::AbstractLoss< unsigned int, RealVector >::evalDerivative(), exportSparseData(), shark::AbstractClustering< RealVector >::hardMembership(), main(), shark::CSvmDerivative< InputType, CacheType >::modelCSvmParameterDerivative(), shark::WeightedDataBatch< DataBatchType, WeightBatchType >::operator[](), shark::DataView< shark::Data< LabelType > const >::operator[](), shark::PointSetKernel< InputType >::weightedParameterDerivative(), and shark::CSvmDerivative< InputType, CacheType >::write().

getBatchElement() [2/2]

template<class BatchT >

auto shark::getBatchElement	(	BatchT const &	batch,
		std::size_t	i
	)		-> decltype(BatchTraits<BatchT>::type::get(std::declval<BatchT const&>(),i))

Definition at line 410 of file BatchInterface.h.

importHDF5() [1/4]

template<typename VectorType >

void shark::importHDF5	(	Data< VectorType > &	data,
		const std::string &	fileName,
		const std::string &	datasetName
	)

Import data from a HDF5 file.

Parameters

data	Container storing the loaded data
fileName	The name of HDF5 file to be read from
datasetName	the HDF5 dataset name to access in the HDF5 file

Template Parameters

VectorType

Type of object stored in Shark data container

Definition at line 268 of file HDF5.h.

importHDF5() [2/4]

template<typename VectorType , typename LabelType >

void shark::importHDF5	(	LabeledData< VectorType, LabelType > &	labeledData,
		const std::string &	fileName,
		const std::string &	data,
		const std::string &	label
	)

Import data to a LabeledData object from a HDF5 file.

Parameters

labeledData	Container storing the loaded data
fileName	The name of HDF5 file to be read from
data	the HDF5 dataset name for data
label	the HDF5 dataset name for label

Template Parameters

VectorType	Type of object stored in Shark data container
LableType	Type of label

Definition at line 294 of file HDF5.h.

importHDF5() [3/4]

template<typename VectorType >

void shark::importHDF5	(	Data< VectorType > &	data,
		const std::string &	fileName,
		const std::vector< std::string > &	cscDatasetName
	)

Import data from HDF5 dataset of compressed sparse column format.

Parameters

data	Container storing the loaded data
fileName	The name of HDF5 file to be read from
cscDatasetName	the CSC dataset names used to construct a matrix

Template Parameters

VectorType

Type of object stored in Shark data container

Definition at line 317 of file HDF5.h.

importHDF5() [4/4]

template<typename VectorType , typename LabelType >

void shark::importHDF5	(	LabeledData< VectorType, LabelType > &	labeledData,
		const std::string &	fileName,
		const std::vector< std::string > &	cscDatasetName,
		const std::string &	label
	)

Import data from HDF5 dataset of compressed sparse column format.

Parameters

labeledData	Container storing the loaded data
fileName	The name of HDF5 file to be read from
cscDatasetName	the CSC dataset names used to construct a matrix
label	the HDF5 dataset name for label

Template Parameters

VectorType	Type of object stored in Shark data container
LabelType	Type of label

Definition at line 343 of file HDF5.h.

importPGM()

template<class T >

void shark::importPGM	(	std::string const &	fileName,
		T &	data,
		std::size_t &	sx,
		std::size_t &	sy
	)

Import a PGM image from file.

Parameters

fileName	The file to read from
data	Linear object for storing image
sx	Width of imported image
sy	Height of imported image

Definition at line 103 of file Pgm.h.

Referenced by importPGMSet().

importPGMSet()

template<class T >

void shark::importPGMSet	(	std::string const &	p,
		Data< T > &	set
	)

Import PGM images scanning a directory recursively.

All images are required to have the same size. the shape of the images is stored in set.shape()

Parameters

p	Directory
set	Set storing images

Definition at line 222 of file Pgm.h.

References createDataFromRange(), importPGM(), and SHARKEXCEPTION.

Referenced by main().

inputDimension() [1/2]

template<class DatasetType >

std::size_t shark::inputDimension

(

DataView< DatasetType > const &

view

)

Return the input dimensionality of the labeled dataset represented by the view.

Definition at line 333 of file DataView.h.

References shark::DataView< DatasetType >::dataset(), and inputDimension().

inputDimension() [2/2]

template<class InputType , class LabelType >

std::size_t shark::inputDimension

(

WeightedLabeledData< InputType, LabelType > const &

dataset

)

Return the input dimensionality of a weighted labeled dataset.

Definition at line 713 of file WeightedDataset.h.

References dataDimension(), and shark::WeightedLabeledData< InputT, LabelT >::inputs().

kMeans() [1/3]

SHARK_EXPORT_SYMBOL std::size_t shark::kMeans	(	Data< RealVector > const &	data,
		std::size_t	k,
		Centroids &	centroids,
		std::size_t	maxIterations = 0
	)

The k-means clustering algorithm.

The k-means algorithm takes vector-valued data \( \{x_1, \dots, x_n\} \subset \mathbb R^d \) and splits it into k clusters, based on centroids \( \{c_1, \dots, c_k\} \). The result is stored in a Centroids object that can be used to construct clustering models.

This implementation starts the search with the given centroids, in case the provided centroids object (third parameter) contains a set of k centroids. Otherwise the search starts from the first k data points.

Note that the data set needs to include at least k data points for k-means to work. This is because the current implementation does not allow for empty clusters.

Parameters

data	vector-valued data to be clustered
k	number of clusters
centroids	centroids input/output
maxIterations	maximum number of k-means iterations; 0: unlimited

Returns

number of k-means iterations

Referenced by main().

kMeans() [2/3]

SHARK_EXPORT_SYMBOL std::size_t shark::kMeans	(	Data< RealVector > const &	data,
		RBFLayer &	model,
		std::size_t	maxIterations = 0
	)

The k-means clustering algorithm for initializing an RBF Layer.

The k-means algorithm takes vector-valued data \( \{x_1, \dots, x_n\} \subset \mathbb R^d \) and splits it into k clusters, based on centroids \( \{c_1, \dots, c_k\} \). The result is stored in a RBFLayer object that can be used to construct clustering models.

This is just an alternative frontend to the version using Centroids. it creates a centroid object, with as many clusters as are outputs in the RBFLayer and copies the result into the model.

Note that the data set needs to include at least k data points for k-means to work. This is because the current implementation does not allow for empty clusters.

Parameters

data	vector-valued data to be clustered
model	RBFLayer input/output
maxIterations	maximum number of k-means iterations; 0: unlimited

Returns

number of k-means iterations

kMeans() [3/3]

template<class InputType >

KernelExpansion<InputType> shark::kMeans	(	Data< InputType > const &	dataset,
		std::size_t	k,
		AbstractKernelFunction< InputType > &	kernel,
		std::size_t	maxIterations = 0
	)

The kernel k-means clustering algorithm.

The kernel k-means algorithm takes data \( \{x_1, \dots, x_n\} \) and splits it into k clusters, based on centroids \( \{c_1, \dots, c_k\} \). The centroids are elements of the reproducing kernel Hilbert space (RHKS) induced by the kernel function. They are functions, represented as the components of a KernelExpansion object. I.e., given a data point x, the kernel expansion returns a k-dimensional vector f(x), which is the evaluation of the centroid functions on x. The value of the centroid function represents the inner product of the centroid with the kernel-induced feature vector of x (embedding of x into the RKHS). The distance of x from the centroid \( c_i \) is computes as the kernel-induced distance \( \sqrt{ kernel(x, x) + kernel(c_i, c_i) - 2 kernel(x, c_i) } \). For the Gaussian kernel (and other normalized kernels) is simplifies to \( \sqrt{ 2 - 2 kernel(x, c_i) } \). Hence, larger function values indicate smaller distance to the centroid.

Note that the data set needs to include at least k data points for k-means to work. This is because the current implementation does not allow for empty clusters.

Parameters

dataset	vector-valued data to be clustered
k	number of clusters
kernel	kernel function object
maxIterations	maximum number of k-means iterations; 0: unlimited

Returns

centroids (represented as functions, see description)

Definition at line 141 of file KMeans.h.

labelDimension() [1/2]

template<class DatasetType >

std::size_t shark::labelDimension

(

DataView< DatasetType > const &

view

)

Return the label dimensionality of the labeled dataset represented by the view.

Definition at line 338 of file DataView.h.

References shark::DataView< DatasetType >::dataset(), and labelDimension().

labelDimension() [2/2]

template<class InputType , class LabelType >

std::size_t shark::labelDimension

(

WeightedLabeledData< InputType, LabelType > const &

dataset

)

Return the label/output dimensionality of a labeled dataset.

Definition at line 719 of file WeightedDataset.h.

References dataDimension(), and shark::WeightedLabeledData< InputT, LabelT >::labels().

logPartitionFunction()

template<class RBMType >

double shark::logPartitionFunction	(	RBMType const &	rbm,
		double	beta = 1.0
	)

Calculates the value of the partition function $Z$.

Only useful for small input and theoretical analysis

Parameters

rbm	the RBM for which to calculate the function
beta	the inverse temperature of the RBM. default is 1

Returns

the value of the partition function $Z*e^(-constant)$

Definition at line 48 of file analytics.h.

References beta.

Referenced by negativeLogLikelihood().

makeKeyValuePair()

template<class Key , class Value >

KeyValuePair<Key,Value> shark::makeKeyValuePair	(	Key const &	key,
		Value const &	value
	)

Creates a KeyValuePair.

Definition at line 96 of file KeyValuePair.h.

References shark::KeyValuePair< Key, Value >::key, and shark::KeyValuePair< Key, Value >::value.

makeResultSet()

template<typename T , typename U >

ResultSet<T,U> shark::makeResultSet	(	T const &	t,
		U const &	u
	)

Generates a typed solution given the search point and the corresponding objective function value.

Parameters

[in]	t	The search point.
[in]	u	The objective function value.

Returns

A ResultSet containing the supplied search point and objective function value.

Definition at line 71 of file ResultSets.h.

mean() [1/2]

template<class VectorType >

VectorType shark::mean

(

Data< VectorType > const &

data

)

Calculates the mean vector of the input vectors.

Referenced by createData(), shark::ZDT6::eval(), shark::TruncatedExponentialLayer::expectedPhiValue(), shark::random::gauss(), getSamples(), shark::McPegasos< VectorType >::lossGradientADS(), shark::McPegasos< VectorType >::lossGradientATS(), mean(), shark::FisherLDA::setWhitening(), shark::statistics::Variance::statistics(), shark::NormalizeComponentsZCA::train(), shark::NormalizeComponentsUnitVariance< DataType >::train(), shark::RegularizationNetworkTrainer< InputType >::train(), shark::QpMcLinearATS< InputT >::updateWeightVectors(), shark::QpMcLinearATM< InputT >::updateWeightVectors(), and shark::QpMcLinearReinforced< InputT >::updateWeightVectors().

mean() [2/2]

template<class VectorType >

VectorType shark::mean

(

UnlabeledData< VectorType > const &

data

)

Definition at line 69 of file Statistics.h.

References covariance(), mean(), and variance().

meanvar() [1/2]

template<class Vec1T , class Vec2T , class Vec3T , class Device >

void shark::meanvar	(	Data< Vec1T > const &	data,
		blas::vector_container< Vec2T, Device > &	mean,
		blas::vector_container< Vec3T, Device > &	variance
	)

Calculates the mean and variance values of the input data.

Referenced by main(), shark::NormalizeComponentsZCA::train(), and shark::NormalizeComponentsUnitVariance< DataType >::train().

meanvar() [2/2]

template<class Vec1T , class Vec2T , class MatT , class Device >

void shark::meanvar	(	Data< Vec1T > const &	data,
		blas::vector_container< Vec2T, Device > &	mean,
		blas::matrix_container< MatT, Device > &	variance
	)

Calculates the mean, variance and covariance values of the input data.

median_element() [1/2]

template<class Range >

boost::range_iterator<Range>::type shark::median_element

(

Range &

range

)

Returns the iterator to the median element. after this call, the range is partially ordered.

After the call, all elements left of the median element are guaranteed to be <= median and all element on the right are >= median.

Definition at line 80 of file functional.h.

Referenced by median_element(), and partitionEqually().

median_element() [2/2]

template<class Range >

boost::range_iterator<Range>::type shark::median_element

(

Range const &

rangeAdaptor

)

Definition at line 91 of file functional.h.

References median_element().

negativeLogLikelihood()

template<class RBMType >

double shark::negativeLogLikelihood	(	RBMType const &	rbm,
		UnlabeledData< RealVector > const &	inputs,
		double	beta = 1.0
	)

Estimates the negative log-likelihood of a set of input vectors under the models distribution.

Only useful for small input and theoretical analysis

Parameters

rbm	the Restricted Boltzmann machine for which the negative log likelihood of the data is to be calculated
inputs	the input vectors
beta	the inverse temperature of the RBM. default is 1

Returns

the log-likelihood

Definition at line 102 of file analytics.h.

References beta, logPartitionFunction(), and negativeLogLikelihoodFromLogPartition().

Referenced by shark::ExactGradient< RBMType >::eval(), and main().

negativeLogLikelihoodFromLogPartition()

template<class RBMType >

double shark::negativeLogLikelihoodFromLogPartition	(	RBMType const &	rbm,
		UnlabeledData< RealVector > const &	inputs,
		double	logPartition,
		double	beta = 1.0
	)

Estimates the negative log-likelihood of a set of input vectors under the models distribution using the partition function.

Only useful for small input and theoretical analysis

Parameters

rbm	the Restricted Boltzmann machine for which the negative log likelihood of the data is to be calculated
inputs	the input vectors
logPartition	the logarithmic value of the partition function of the RBM.
beta	the inverse temperature of the RBM. default is 1

Returns

the log-likelihood

Definition at line 79 of file analytics.h.

References shark::Data< Type >::batches().

Referenced by negativeLogLikelihood().

nonDominatedSort() [1/2]

template<class PointRange , class RankRange >

void shark::nonDominatedSort	(	PointRange const &	points,
		RankRange &	ranks
	)

Frontend for non-dominated sorting algorithms.

Assembles subsets/fronts of mutually non-dominated individuals. Afterwards every individual is assigned a rank by pop[i].rank() = frontIndex. The front of non-dominated points has the value 1.

Depending on dimensionality m and number of points n, either the fastNonDominatedSort algorithm with O(n^2 m) or the dcNonDominatedSort alforithm with complexity O(n log(n)^m) is called.

Definition at line 47 of file NonDominatedSort.h.

References dcNonDominatedSort(), fastNonDominatedSort(), and SIZE_CHECK.

Referenced by nonDominatedSort(), and shark::IndicatorBasedSelection< NSGA3Indicator >::operator()().

nonDominatedSort() [2/2]

template<class PointRange , class RankRange >

void shark::nonDominatedSort	(	PointRange const &	points,
		RankRange const &	ranks
	)

Definition at line 68 of file NonDominatedSort.h.

References nonDominatedSort(), and ranks().

numberOfClasses() [1/3]

template<class DatasetType >

std::size_t shark::numberOfClasses

(

DataView< DatasetType > const &

view

)

Return the number of classes (size of the label vector) of a classification dataset with RealVector label encoding.

Definition at line 327 of file DataView.h.

References shark::DataView< DatasetType >::dataset(), and numberOfClasses().

numberOfClasses() [2/3]

std::size_t shark::numberOfClasses ( WeightedUnlabeledData< unsigned int > const &  labels )

inline

Definition at line 696 of file WeightedDataset.h.

References numberOfClasses().

numberOfClasses() [3/3]

template<class InputType >

std::size_t shark::numberOfClasses

(

WeightedLabeledData< InputType, unsigned int > const &

dataset

)

Return the number of classes (highest label value +1) of a classification dataset with unsigned int label encoding.

Definition at line 724 of file WeightedDataset.h.

References shark::WeightedLabeledData< InputT, LabelT >::labels(), and numberOfClasses().

operator!=()

bool shark::operator!= ( Shape const &  shape1, Shape const &  shape2  )

inline

Definition at line 113 of file Shape.h.

operator<<() [1/5]

template<class SearchPoint , class Result >

std::ostream& shark::operator<<	(	std::ostream &	out,
		ResultSet< SearchPoint, Result > const &	solution
	)

Definition at line 75 of file ResultSets.h.

References shark::ResultSet< SearchPointT, ResultT >::value.

operator<<() [2/5]

template<class E , class T >

std::basic_ostream<E, T>& shark::operator<<	(	std::basic_ostream< E, T > &	os,
		Shape const &	shape
	)

Definition at line 118 of file Shape.h.

operator<<() [3/5]

template<class SearchPoint >

std::ostream& shark::operator<<	(	std::ostream &	out,
		ValidatedSingleObjectiveResultSet< SearchPoint > const &	solution
	)

Definition at line 131 of file ResultSets.h.

operator<<() [4/5]

template<class T >

std::ostream& shark::operator<<	(	std::ostream &	stream,
		const WeightedUnlabeledData< T > &	d
	)

brief Outstream of elements for weighted data.

Definition at line 531 of file WeightedDataset.h.

operator<<() [5/5]

template<class T , class U >

std::ostream& shark::operator<<	(	std::ostream &	stream,
		const WeightedLabeledData< T, U > &	d
	)

brief Outstream of elements for weighted labeled data.

Definition at line 688 of file WeightedDataset.h.

operator==()

bool shark::operator== ( Shape const &  shape1, Shape const &  shape2  )

inline

Definition at line 102 of file Shape.h.

References shark::Shape::size().

operator>>() [1/2]

template<class VectorType >

ConcatenatedModel<VectorType> shark::operator>>	(	AbstractModel< VectorType, VectorType, VectorType > &	firstModel,
		AbstractModel< VectorType, VectorType, VectorType > &	secondModel
	)

Connects two AbstractModels so that the output of the first model is the input of the second.

Definition at line 308 of file ConcatenatedModel.h.

References shark::ConcatenatedModel< VectorType >::add().

operator>>() [2/2]

template<class VectorType >

ConcatenatedModel<VectorType> shark::operator>>	(	ConcatenatedModel< VectorType > const &	firstModel,
		AbstractModel< VectorType, VectorType, VectorType > &	secondModel
	)

Definition at line 319 of file ConcatenatedModel.h.

References shark::ConcatenatedModel< VectorType >::add().

partial_shuffle() [1/2]

template<class RandomAccessIterator , class Rng >

void shark::partial_shuffle	(	RandomAccessIterator	begin,
		RandomAccessIterator	middle,
		RandomAccessIterator	end,
		Rng &	rng
	)

random_shuffle algorithm which stops after acquiring the random subsequence for [begin,middle)

Definition at line 55 of file functional.h.

References shuffle().

Referenced by partial_shuffle(), randomSubBatch(), and randomSubset().

partial_shuffle() [2/2]

template<class RandomAccessIterator >

void shark::partial_shuffle	(	RandomAccessIterator	begin,
		RandomAccessIterator	middle,
		RandomAccessIterator	end
	)

random_shuffle algorithm which stops after acquiring the random subsequence for [begin,middle)

Definition at line 71 of file functional.h.

References shark::random::globalRng, and partial_shuffle().

partitionEqually() [1/2]

template<class Range >

boost::range_iterator<Range>::type shark::partitionEqually

(

Range &

range

)

Partitions a range in two parts as equal in size as possible.

The Algorithm partitions the range and returns the splitpoint. The elements in the range are ordered such that all elements in [begin,splitpoint) < [splitpoint,end). This partition is done such that the ranges are as equally sized as possible. It is guaranteed that the left range is not empty. However, if the range consists only of equal elements, the return value will be the end iterator indicating that there is no split possible. The whole algorithm runs in linear time by iterating 2 times over the sequence.

Definition at line 105 of file functional.h.

References median_element().

Referenced by partitionEqually(), and shark::BinaryTree< VectorType >::splitList().

partitionEqually() [2/2]

template<class Range >

boost::range_iterator<Range>::type shark::partitionEqually

(

Range const &

rangeAdaptor

)

Partitions a range in two parts as equal in size as possible and returns it's result.

This the verison for adapted ranges.

Definition at line 133 of file functional.h.

References partitionEqually().

penalizedFitness()

template<class IndividualRange >

auto shark::penalizedFitness

(

IndividualRange &

range

)

-> decltype( boost::adaptors::transform(range,detail::IndividualPenalizedFitnessFunctor()) )

Definition at line 228 of file Individual.h.

References transform().

Referenced by shark::IndicatorBasedSelection< NSGA3Indicator >::indicator(), shark::IndicatorBasedSelection< NSGA3Indicator >::operator()(), and shark::PopulationBasedStepSizeAdaptation::update().

preferenceAdjustedUnitVectors()

RealMatrix shark::preferenceAdjustedUnitVectors	(	std::size_t const	n,
		std::size_t const	sum,
		std::vector< Preference > const &	preferences
	)

Return a set of evenly spaced n-dimensional points on the unit sphere clustered around the specified preference points.

Referenced by shark::NSGA3Indicator::init().

preferenceAdjustedWeightVectors()

RealMatrix shark::preferenceAdjustedWeightVectors	(	std::size_t const	n,
		std::size_t const	sum,
		std::vector< Preference > const &	preferences
	)

Return a set of of evenly spaced n-dimensional points on the "unit simplex" clustered around the specified preference points.

randomSubBatch()

template<class DatasetType >

DataView<DatasetType>::batch_type shark::randomSubBatch	(	DataView< DatasetType > const &	view,
		std::size_t	size
	)

Creates a random batch of a given size.

Parameters

view	the view from which the batch is to be created
size	the size of the batch

Definition at line 285 of file DataView.h.

References partial_shuffle(), shark::DataView< DatasetType >::size(), and subBatch().

randomSubset()

template<class DatasetType >

DataView<DatasetType> shark::randomSubset	(	DataView< DatasetType > const &	view,
		std::size_t	size
	)

creates a random subset of a DataView with given size

Parameters

view	the view for which the subset is to be created
size	the size of the subset

Definition at line 257 of file DataView.h.

References partial_shuffle(), shark::DataView< DatasetType >::size(), and subset().

Referenced by main().

ranks()

template<class IndividualRange >

auto shark::ranks

(

IndividualRange &

range

)

-> decltype( boost::adaptors::transform(range,detail::IndividualRankFunctor()) )

Definition at line 242 of file Individual.h.

References transform().

Referenced by nonDominatedSort(), and shark::IndicatorBasedSelection< NSGA3Indicator >::operator()().

read_front()

template<typename Stream >

FrontType shark::read_front	(	Stream &	in,
		std::size_t	noObjectives,
		const std::string &	separator = " ",
		std::size_t	headerLines = 0
	)

Definition at line 17 of file AdditiveEpsilonIndicatorMain.cpp.

Referenced by main().

sampleLatticeUniformly()

template<typename Matrix , typename randomType = shark::random::rng_type>

Matrix shark::sampleLatticeUniformly	(	randomType &	rng,
		Matrix const &	matrix,
		std::size_t const	n,
		bool const	keep_corners = true
	)

Definition at line 90 of file Lattice.h.

Referenced by shark::NSGA3Indicator::init().

searchPoint()

template<class IndividualRange >

auto shark::searchPoint

(

IndividualRange &

range

)

-> decltype( boost::adaptors::transform(range,detail::IndividualSearchPointFunctor()) )

Definition at line 250 of file Individual.h.

References transform().

Referenced by shark::LineSearch::init(), main(), shark::VDCMA::step(), and shark::LMCMA::step().

SHARK_VECTOR_MATRIX_TYPEDEFS() [1/2]

shark::SHARK_VECTOR_MATRIX_TYPEDEFS	(	long	double,
		BigReal
	)

SHARK_VECTOR_MATRIX_TYPEDEFS() [2/2]

shark::SHARK_VECTOR_MATRIX_TYPEDEFS	(	bool	,
		Bool
	)

shuffle()

template<class Iterator , class Rng >

void shark::shuffle	(	Iterator	begin,
		Iterator	end,
		Rng &	rng
	)

random_shuffle algorithm which stops after acquiring the random subsequence for [begin,middle)

Definition at line 44 of file functional.h.

References shark::random::discrete(), and swap().

Referenced by createCVBatch(), shark::ContrastiveDivergence< Operator >::evalDerivative(), partial_shuffle(), shark::UnlabeledData< InputType >::shuffle(), and shark::LabeledData< InputType, LabelType >::shuffle().

subBatch()

template<class DatasetType , class IndexRange >

DataView<DatasetType>::batch_type shark::subBatch	(	DataView< DatasetType > const &	view,
		IndexRange const &	indizes
	)

Creates a batch given a set of indices.

Parameters

view	the view from which the batch is to be created
indizes	the set of indizes defining the batch

Definition at line 269 of file DataView.h.

References createBatch(), and subset().

Referenced by shark::RFClassifier< LabelType >::computeFeatureImportances(), createCVFullyIndexed(), createCVIndexed(), and randomSubBatch().

subset()

template<class DatasetType , class IndexRange >

DataView<DatasetType> shark::subset	(	DataView< DatasetType > const &	view,
		IndexRange const &	indizes
	)

creates a subset of a DataView with elements indexed by indices

Parameters

view	the view for which the subset is to be created
indizes	the index of the elements to be stored in the view

Definition at line 247 of file DataView.h.

Referenced by shark::Data< LabelT >::indexedSubset(), main(), randomSubset(), and subBatch().

sumOfWeights() [1/2]

template<class InputType >

double shark::sumOfWeights

(

WeightedUnlabeledData< InputType > const &

dataset

)

Returns the total sum of weights.

Definition at line 736 of file WeightedDataset.h.

sumOfWeights() [2/2]

template<class InputType , class LabelType >

double shark::sumOfWeights

(

WeightedLabeledData< InputType, LabelType > const &

dataset

)

Returns the total sum of weights.

Definition at line 745 of file WeightedDataset.h.

swap() [1/3]

template<class D1 , class W1 , class D2 , class W2 >

void shark::swap	(	WeightedDataPair< D1, W1 > &&	p1,
		WeightedDataPair< D2, W2 > &&	p2
	)

Definition at line 84 of file WeightedDataset.h.

References swap().

swap() [2/3]

template<class K , class V >

void shark::swap	(	KeyValuePair< K, V > &	pair1,
		KeyValuePair< K, V > &	pair2
	)

Swaps the contents of two instances of KeyValuePair.

Definition at line 88 of file KeyValuePair.h.

References shark::KeyValuePair< Key, Value >::key, and shark::KeyValuePair< Key, Value >::value.

Referenced by createCVFullyIndexed(), createCVIndexed(), shark::QpMcBoxDecomp< Matrix >::deactivateExample(), shark::QpMcSimplexDecomp< Matrix >::deactivateExample(), shark::QpMcBoxDecomp< Matrix >::deactivateVariable(), shark::QpMcSimplexDecomp< Matrix >::deactivateVariable(), shark::ConcatenatedModel< VectorType >::eval(), shark::MultiChainApproximator< MarkovChainType >::evalDerivative(), shark::ExampleModifiedKernelMatrix< InputType, CacheType >::flipColumnsAndRows(), shark::BlockMatrix2x2< Matrix >::flipColumnsAndRows(), shark::RegularizedKernelMatrix< InputType, CacheType >::flipColumnsAndRows(), shark::GaussianKernelMatrix< T, CacheType >::flipColumnsAndRows(), shark::KernelMatrix< InputType, CacheType >::flipColumnsAndRows(), shark::ModifiedKernelMatrix< InputType, CacheType >::flipColumnsAndRows(), shark::DifferenceKernelMatrix< InputType, CacheType >::flipColumnsAndRows(), shark::CachedMatrix< Matrix >::flipColumnsAndRows(), shark::GeneralQuadraticProblem< MatrixT >::flipCoordinates(), shark::BoxedSVMProblem< MatrixT >::flipCoordinates(), shark::BoxBasedShrinkingStrategy< BoxConstrainedProblem< Problem > >::flipCoordinates(), shark::BoxConstrainedProblem< Problem >::flipCoordinates(), shark::CSVMProblem< MatrixT >::flipCoordinates(), shark::SvmProblem< Problem >::flipCoordinates(), main(), shark::WS2MaximumGradientCriterion::operator()(), shark::SimulatedBinaryCrossover< RealVector >::operator()(), shark::HMGSelectionCriterion::operator()(), repartitionByClass(), shark::MultiChainApproximator< MarkovChainType >::setData(), shark::BoxBasedShrinkingStrategy< BoxConstrainedProblem< Problem > >::shrink(), shuffle(), swap(), shark::LRUCache< QpFloatType >::swapLineIndices(), shark::ConcatenatedModel< VectorType >::weightedDerivatives(), shark::ConcatenatedModel< VectorType >::weightedInputDerivative(), and shark::ConcatenatedModel< VectorType >::weightedParameterDerivative().

swap() [3/3]

template<class D1 , class W1 , class D2 , class W2 >

void shark::swap	(	WeightedDataBatch< D1, W1 > &	p1,
		WeightedDataBatch< D2, W2 > &	p2
	)

Definition at line 159 of file WeightedDataset.h.

References shark::WeightedDataBatch< DataBatchType, WeightBatchType >::data, swap(), and shark::WeightedDataBatch< DataBatchType, WeightBatchType >::weight.

Referenced by swap().

tchebycheffScalarizer()

double shark::tchebycheffScalarizer	(	RealVector const &	fitness,
		RealVector const &	weights,
		RealVector const &	optimalPointFitness
	)

Definition at line 41 of file Tchebycheff.h.

References w.

toDataset()

template<class T >

DataView<T>::dataset_type shark::toDataset	(	DataView< T > const &	view,
		std::size_t	batchSize = DataView<T>::dataset_type::DefaultBatchSize
	)

Creates a new dataset from a View.

When the elements of a View needs to be processed repeatedly it is often better to use the packed format of the Dataset again, since then the faster batch processing can be used

Parameters

view	the view from which to create the new dataset
batchSize	the size of the batches in the dataset

Definition at line 314 of file DataView.h.

References batchSize(), shark::DataView< DatasetType >::begin(), shark::DataView< DatasetType >::dataset(), shark::DataView< DatasetType >::end(), and shark::DataView< DatasetType >::size().

Referenced by main().

toView()

template<class DatasetType >

DataView<DatasetType> shark::toView

(

DatasetType &

set

)

Creates a View from a dataset.

This is just a helper function to omit the actual type of the view

Parameters

set	the dataset from which to create the view

Definition at line 301 of file DataView.h.

unitVectorsOnLattice()

RealMatrix shark::unitVectorsOnLattice	(	std::size_t const	n,
		std::size_t const	sum
	)

Return a set of evenly spaced n-dimensional points on the unit sphere.

Referenced by shark::NSGA3Indicator::init().

unpenalizedFitness()

template<class IndividualRange >

auto shark::unpenalizedFitness

(

IndividualRange &

range

)

-> decltype( boost::adaptors::transform(range,detail::IndividualUnpenalizedFitnessFunctor()) )

Definition at line 235 of file Individual.h.

References transform().

Referenced by shark::IndicatorBasedSelection< NSGA3Indicator >::indicator(), and shark::ReferenceVectorAdaptation< shark::Individual >::operator()().

variance()

template<class VectorType >

VectorType shark::variance

(

Data< VectorType > const &

data

)

Calculates the variance vector of the input vectors.

Referenced by shark::random::gauss(), getSamples(), mean(), shark::statistics::Variance::statistics(), and shark::NormalizeComponentsUnitVariance< DataType >::train().

weightLattice()

RealMatrix shark::weightLattice	(	std::size_t const	n,
		std::size_t const	sum
	)

Returns a set of evenly spaced n-dimensional points on the "unit simplex".