Training of an SVM for ranking. More...
#include <shark/Algorithms/Trainers/RankingSvmTrainer.h>
Public Types | |
typedef CacheType | QpFloatType |
Convenience typedefs: this and many of the below typedefs build on the class template type CacheType. Simply changing that one template parameter CacheType thus allows to flexibly switch between using float or double as type for caching the kernel values. The default is float, offering sufficient accuracy in the vast majority of cases, at a memory cost of only four bytes. However, the template parameter makes it easy to use double instead, (e.g., in case high accuracy training is needed). More... | |
typedef AbstractKernelFunction< InputType > | KernelType |
Public Types inherited from shark::AbstractSvmTrainer< InputType, unsigned int, KernelExpansion< InputType > > | |
typedef AbstractKernelFunction< InputType > | KernelType |
Public Types inherited from shark::AbstractTrainer< Model, LabelTypeT > | |
typedef Model | ModelType |
typedef ModelType::InputType | InputType |
typedef LabelTypeT | LabelType |
typedef LabeledData< InputType, LabelType > | DatasetType |
Public Types inherited from shark::IParameterizable<> | |
typedef RealVector | ParameterVectorType |
Public Member Functions | |
RankingSvmTrainer (KernelType *kernel, double C, bool unconstrained=false) | |
std::string | name () const |
From INameable: return the class name. More... | |
void | train (KernelExpansion< InputType > &function, Data< InputType > const &dataset) |
Train the ranking SVM. More... | |
void | train (KernelExpansion< InputType > &function, LabeledData< InputType, unsigned int > const &dataset) |
Train the ranking SVM. More... | |
void | train (KernelExpansion< InputType > &function, Data< InputType > const &dataset, std::vector< std::pair< std::size_t, std::size_t >> const &pairs) |
Train the ranking SVM. More... | |
Public Member Functions inherited from shark::AbstractSvmTrainer< InputType, unsigned int, KernelExpansion< InputType > > | |
AbstractSvmTrainer (KernelType *kernel, double C, bool offset, bool unconstrained=false) | |
AbstractSvmTrainer (KernelType *kernel, double negativeC, double positiveC, bool offset, bool unconstrained=false) | |
double | C () const |
Return the value of the regularization parameter C. More... | |
void | setC (double C) |
Set the value of the regularization parameter C. More... | |
RealVector const & | regularizationParameters () const |
void | setRegularizationParameters (RealVector const ®ularizers) |
Set the value of the regularization parameter C. More... | |
KernelType * | kernel () |
KernelType const * | kernel () const |
void | setKernel (KernelType *kernel) |
bool | isUnconstrained () const |
bool | trainOffset () const |
std::size_t | cacheSize () const |
void | setCacheSize (std::size_t size) |
RealVector | parameterVector () const |
get the hyper-parameter vector More... | |
void | setParameterVector (RealVector const &newParameters) |
set the vector of hyper-parameters More... | |
size_t | numberOfParameters () const |
return the number of hyper-parameters More... | |
Public Member Functions inherited from shark::AbstractTrainer< Model, LabelTypeT > | |
virtual void | train (ModelType &model, DatasetType const &dataset)=0 |
Core of the Trainer interface. More... | |
Public Member Functions inherited from shark::INameable | |
virtual | ~INameable () |
Public Member Functions inherited from shark::ISerializable | |
virtual | ~ISerializable () |
Virtual d'tor. More... | |
virtual void | read (InArchive &archive) |
Read the component from the supplied archive. More... | |
virtual void | write (OutArchive &archive) const |
Write the component to the supplied archive. More... | |
void | load (InArchive &archive, unsigned int version) |
Versioned loading of components, calls read(...). More... | |
void | save (OutArchive &archive, unsigned int version) const |
Versioned storing of components, calls write(...). More... | |
BOOST_SERIALIZATION_SPLIT_MEMBER () | |
Public Member Functions inherited from shark::QpConfig | |
QpConfig (bool precomputedFlag=false, bool sparsifyFlag=true) | |
Constructor. More... | |
QpStoppingCondition & | stoppingCondition () |
Read/write access to the stopping condition. More... | |
QpStoppingCondition const & | stoppingCondition () const |
Read access to the stopping condition. More... | |
QpSolutionProperties & | solutionProperties () |
Access to the solution properties. More... | |
bool & | precomputeKernel () |
Flag for using a precomputed kernel matrix. More... | |
bool const & | precomputeKernel () const |
Flag for using a precomputed kernel matrix. More... | |
bool & | sparsify () |
Flag for sparsifying the model after training. More... | |
bool const & | sparsify () const |
Flag for sparsifying the model after training. More... | |
bool & | shrinking () |
Flag for shrinking in the decomposition solver. More... | |
bool const & | shrinking () const |
Flag for shrinking in the decomposition solver. More... | |
bool & | s2do () |
Flag for S2DO (instead of SMO) More... | |
bool const & | s2do () const |
Flag for S2DO (instead of SMO) More... | |
unsigned int & | verbosity () |
Verbosity level of the solver. More... | |
unsigned int const & | verbosity () const |
Verbosity level of the solver. More... | |
unsigned long long const & | accessCount () const |
Number of kernel accesses. More... | |
void | setMinAccuracy (double a) |
void | setMaxIterations (unsigned long long i) |
void | setTargetValue (double v) |
void | setMaxSeconds (double s) |
Public Member Functions inherited from shark::IParameterizable<> | |
virtual | ~IParameterizable () |
Additional Inherited Members | |
Protected Attributes inherited from shark::AbstractSvmTrainer< InputType, unsigned int, KernelExpansion< InputType > > | |
KernelType * | m_kernel |
RealVector | m_regularizers |
Vector of regularization parameters. More... | |
bool | m_trainOffset |
bool | m_unconstrained |
Is log(C) stored internally as a parameter instead of C? If yes, then we get rid of the constraint C > 0 on the level of the parameter interface. More... | |
std::size_t | m_cacheSize |
Number of values in the kernel cache. The size of the cache in bytes is the size of one entry (4 for float, 8 for double) times this number. More... | |
Protected Attributes inherited from shark::QpConfig | |
QpStoppingCondition | m_stoppingcondition |
conditions for when to stop the QP solver More... | |
QpSolutionProperties | m_solutionproperties |
properties of the approximate solution found by the solver More... | |
bool | m_precomputedKernelMatrix |
should the solver use a precomputed kernel matrix? More... | |
bool | m_sparsify |
should the trainer sparsify the model after training? More... | |
bool | m_shrinking |
should shrinking be used? More... | |
bool | m_s2do |
should S2DO be used instead of SMO? More... | |
unsigned int | m_verbosity |
verbosity level (currently unused) More... | |
unsigned long long | m_accessCount |
kernel access count More... | |
Training of an SVM for ranking.
A ranking SVM trains a function (linear or linear in a kernel induced feature space, RKHS) with the aim that the function values are consistent with given pairwise rankings. I.e., given are pairs (a, b) of points, and the task of SVM training is to find a function f such that f(a) < f(b). More exactly, the hard margin ranking SVM aims for f(b) - f(a) >= 1 while minimizing the squared RKHS norm of f. The soft-margin SVM relates the constraint analogous to a standard C-SVM.
The trained model is a real-valued function. To predict the ranking of a pair of points the function is applied to both points. The one with smaller function value is ranked as being "smaller", i.e., if f is the trained model and a and b are data points, then the following code computes the ranking:
bool a_better_than_b = (f(a) < f(b));
Definition at line 71 of file RankingSvmTrainer.h.
typedef AbstractKernelFunction<InputType> shark::RankingSvmTrainer< InputType, CacheType >::KernelType |
Definition at line 87 of file RankingSvmTrainer.h.
typedef CacheType shark::RankingSvmTrainer< InputType, CacheType >::QpFloatType |
Convenience typedefs: this and many of the below typedefs build on the class template type CacheType. Simply changing that one template parameter CacheType thus allows to flexibly switch between using float or double as type for caching the kernel values. The default is float, offering sufficient accuracy in the vast majority of cases, at a memory cost of only four bytes. However, the template parameter makes it easy to use double instead, (e.g., in case high accuracy training is needed).
Definition at line 85 of file RankingSvmTrainer.h.
|
inline |
Constructor
kernel | kernel function to use for training and prediction |
C | regularization parameter - always the 'true' value of C, even when unconstrained is set |
unconstrained | when a C-value is given via setParameter, should it be piped through the exp-function before using it in the solver? |
Definition at line 93 of file RankingSvmTrainer.h.
|
inlinevirtual |
From INameable: return the class name.
Reimplemented from shark::INameable.
Definition at line 98 of file RankingSvmTrainer.h.
|
inline |
Train the ranking SVM.
This variant of the train function assumes that all pairs of points should be ranked according to the order they appear in the data set.
Definition at line 106 of file RankingSvmTrainer.h.
References shark::Data< Type >::numberOfElements().
Referenced by shark::RankingSvmTrainer< InputType, CacheType >::train().
|
inline |
Train the ranking SVM.
This variant of the train function uses integer labels to define pairwise rankings. It is trained on all pairs of data points with different label, aiming for a smaller function value for the point with smaller label.
Definition at line 125 of file RankingSvmTrainer.h.
References shark::Data< Type >::elements(), shark::LabeledData< InputT, LabelT >::inputs(), shark::LabeledData< InputT, LabelT >::labels(), and shark::RankingSvmTrainer< InputType, CacheType >::train().
|
inline |
Train the ranking SVM.
This variant of the train function works with explicitly given pairs of data points. Each pair is identified by the indices of the training points in the data set.
Definition at line 147 of file RankingSvmTrainer.h.
References shark::GeneralQuadraticProblem< MatrixT >::boxMax, shark::GeneralQuadraticProblem< MatrixT >::boxMin, shark::AbstractSvmTrainer< InputType, unsigned int, KernelExpansion< InputType > >::C(), shark::GeneralQuadraticProblem< MatrixT >::dimensions(), shark::AbstractSvmTrainer< InputType, unsigned int, KernelExpansion< InputType > >::kernel(), shark::GeneralQuadraticProblem< MatrixT >::linear, shark::AbstractSvmTrainer< InputType, unsigned int, KernelExpansion< InputType > >::m_kernel, shark::QpConfig::m_shrinking, shark::Data< Type >::numberOfElements(), shark::QpConfig::precomputeKernel(), SIZE_CHECK, shark::QpConfig::solutionProperties(), shark::QpSolver< Problem, SelectionStrategy >::solve(), and shark::QpConfig::stoppingCondition().