Detailed Description

Abstract base class that provides an interface for performing kernel two-sample test using Maximum Mean Discrepancy (MMD) as the test statistic. The MMD is the distance of two probability distributions \(p\) and \(q\) in a RKHS (see [1] for formal description).

\[ \text{MMD}[\mathcal{F},p,q]^2=||\mu_p - \mu_q||^2_\mathcal{F}= \textbf{E}_{x,x'}\left[ k(x,x')\right] -2\textbf{E}_{x,y}\left[ k(x,y)\right] +\textbf{E}_{y,y'}\left[ k(y,y')\right] \]

where \(x,x'\sim p\) and \(y,y'\sim q\).

Given two sets of samples \(\{x_i\}_{i=1}^{n_x}\sim p\) and \(\{y_i\}_{i=1}^{n_y}\sim q\), \(n_x+n_y=n\), the unbiased estimate of the above statistic is computed as

\[ \hat{\eta}_{k,U}=\frac{1}{n_x(n_x-1)}\sum_{i=1}^{n_x}\sum_{j\neq i} k(x_i,x_j)+\frac{1}{n_y(n_y-1)}\sum_{i=1}^{n_y}\sum_{j\neq i}k(y_i,y_j) -\frac{2}{n_xn_y}\sum_{i=1}^{n_x}\sum_{j=1}^{n_y}k(x_i,y_j) \]

A biased version is

\[ \hat{\eta}_{k,V}=\frac{1}{n_x^2}\sum_{i=1}^{n_x}\sum_{j=1}^{n_x} k(x_i,x_j)+\frac{1}{n_y^2}\sum_{i=1}^{n_y}\sum_{j=1}^{n_y}k(y_i,y_j) -\frac{2}{n_xn_y}\sum_{i=1}^{n_x}\sum_{j=1}^{n_y}k(x_i,y_j) \]

When \(n_x=n_y=\frac{n}{2}\), an incomplete version can also be computed as the following

\[ \hat{\eta}_{k,U^-}=\frac{1}{\frac{n}{2}(\frac{n}{2}-1)}\sum_{i\neq j} h(z_i,z_j) \]

where for each pair \(z=(x,y)\), \(h(z,z')=k(x,x')+k(y,y')-k(x,y')- k(x',y)\).

The type (biased/unbiased/incomplete) can be selected via set_statistic_type() via the enum values from EStatisticType, ST_BIASED, ST_UNBIASED and ST_INCOMPLETE, respectively. The estimate returned by compute_statistic() is \(\frac{n_xn_y}{n_x+n_y}\hat{\eta}_k\).

This class provides an interface for adding multiple kernels and then selecting the best kernel based on specified strategies. To know more in details about various learning algorithms for optimal kernel selection, please refer to [2].

Along with the statistic comes a method to compute a p-value based on different methods. Permutation test is possible. If unsure which one to use, sampling with 250 permutation iterations usually always is correct.

To choose, use set_null_approximation_method() and choose from.

NAM_MMD2_SPECTRUM: For a fast, consistent test based on the spectrum of the kernel matrix, as described in [2]. Only supported if Eigen3 is installed. Only applicable for CQuadraticTimeMMD.

NAM_MMD2_GAMMA: for a very fast, but not consistent test based on moment matching of a Gamma distribution, as described in [2]. Only applicable for CQuadraticTimeMMD.

NAM_PERMUTATION: For permuting available samples to sample null-distribution

[1]: Gretton, A., Borgwardt, K. M., Rasch, M. J., Schoelkopf, B., & Smola, A. (2012). A Kernel Two-Sample Test. Journal of Machine Learning Research, 13, 671-721.

[2] Arthur Gretton, Bharath K. Sriperumbudur, Dino Sejdinovic, Heiko Strathmann, Sivaraman Balakrishnan, Massimiliano Pontil, Kenji Fukumizu: Optimal kernel choice for large-scale two-sample tests. NIPS 2012: 1214-1222.

Definition at line 120 of file MMD.h.

Inheritance diagram for CMMD:

[legend]

Classes
struct	Self

Public Types
typedef rxcpp::subjects::subject< ObservedValue >	SGSubject

typedef rxcpp::observable< ObservedValue, rxcpp::dynamic_observable< ObservedValue > >	SGObservable

typedef rxcpp::subscriber< ObservedValue, rxcpp::observer< ObservedValue, void, void, void, void > >	SGSubscriber

Public Member Functions
	CMMD ()

	CMMD (CFeatures samples_from_p, CFeatures samples_from_q)

virtual	~CMMD ()

void	set_kernel_selection_strategy (EKernelSelectionMethod method, bool weighted=false)

void	set_kernel_selection_strategy (EKernelSelectionMethod method, index_t num_runs, index_t num_folds, float64_t alpha)

void	add_kernel (CKernel *kernel)

virtual void	select_kernel ()

CKernelSelectionStrategy const *	get_kernel_selection_strategy () const

virtual float64_t	compute_statistic ()=0

virtual SGVector< float64_t >	sample_null ()=0

void	cleanup ()

void	set_num_null_samples (index_t null_samples)

index_t	get_num_null_samples () const

void	set_statistic_type (EStatisticType stype)

EStatisticType	get_statistic_type () const

void	set_null_approximation_method (ENullApproximationMethod nmethod)

ENullApproximationMethod	get_null_approximation_method () const

virtual const char *	get_name () const

virtual void	set_kernel (CKernel *kernel)

CKernel *	get_kernel () const

virtual void	set_p (CFeatures *samples_from_p)

CFeatures *	get_p () const

virtual void	set_q (CFeatures *samples_from_q)

CFeatures *	get_q () const

void	set_num_samples_p (index_t num_samples_from_p)

const index_t	get_num_samples_p () const

void	set_num_samples_q (index_t num_samples_from_q)

const index_t	get_num_samples_q () const

CCustomDistance *	compute_distance (CDistance *distance)

CCustomDistance *	compute_joint_distance (CDistance *distance)

void	set_train_test_mode (bool on)

void	set_train_test_ratio (float64_t ratio)

virtual float64_t	compute_p_value (float64_t statistic)

virtual float64_t	compute_threshold (float64_t alpha)

bool	perform_test (float64_t alpha)

int32_t	ref ()

int32_t	ref_count ()

int32_t	unref ()

virtual CSGObject *	shallow_copy () const

virtual CSGObject *	deep_copy () const

virtual bool	is_generic (EPrimitiveType *generic) const

template<class T >
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

template<>
void	set_generic ()

void	unset_generic ()

virtual void	print_serializable (const char *prefix="")

virtual bool	save_serializable (CSerializableFile file, const char prefix="")

virtual bool	load_serializable (CSerializableFile file, const char prefix="")

void	set_global_io (SGIO *io)

SGIO *	get_global_io ()

void	set_global_parallel (Parallel *parallel)

Parallel *	get_global_parallel ()

void	set_global_version (Version *version)

Version *	get_global_version ()

SGStringList< char >	get_modelsel_names ()

void	print_modsel_params ()

char *	get_modsel_param_descr (const char *param_name)

index_t	get_modsel_param_index (const char *param_name)

void	build_gradient_parameter_dictionary (CMap< TParameter , CSGObject > *dict)

bool	has (const std::string &name) const

template<typename T >
bool	has (const Tag< T > &tag) const

template<typename T , typename U = void>
bool	has (const std::string &name) const

template<typename T >
void	set (const Tag< T > &_tag, const T &value)

template<typename T , typename U = void>
void	set (const std::string &name, const T &value)

template<typename T >
T	get (const Tag< T > &_tag) const

template<typename T , typename U = void>
T	get (const std::string &name) const

SGObservable *	get_parameters_observable ()

void	subscribe_to_parameters (ParameterObserverInterface *obs)

void	list_observable_parameters ()

virtual void	update_parameter_hash ()

virtual bool	parameter_hash_changed ()

virtual bool	equals (CSGObject *other, float64_t accuracy=0.0, bool tolerant=false)

virtual CSGObject *	clone ()

Public Attributes
SGIO *	io

Parallel *	parallel

Version *	version

Parameter *	m_parameters

Parameter *	m_model_selection_parameters

Parameter *	m_gradient_parameters

uint32_t	m_hash

Protected Member Functions
virtual float64_t	normalize_statistic (float64_t statistic) const =0

internal::KernelManager &	get_kernel_mgr ()

const internal::KernelManager &	get_kernel_mgr () const

internal::DataManager &	get_data_mgr ()

const internal::DataManager &	get_data_mgr () const

virtual void	load_serializable_pre () throw (ShogunException)

virtual void	load_serializable_post () throw (ShogunException)

virtual void	save_serializable_pre () throw (ShogunException)

virtual void	save_serializable_post () throw (ShogunException)

template<typename T >
void	register_param (Tag< T > &_tag, const T &value)

template<typename T >
void	register_param (const std::string &name, const T &value)

bool	clone_parameters (CSGObject *other)

void	observe (const ObservedValue value)

void	register_observable_param (const std::string &name, const SG_OBS_VALUE_TYPE type, const std::string &description)

Member Typedef Documentation

◆ SGObservable

typedef rxcpp::observable<ObservedValue, rxcpp::dynamic_observable<ObservedValue> > SGObservable

inherited

Definition at line 130 of file SGObject.h.

◆ SGSubject

typedef rxcpp::subjects::subject<ObservedValue> SGSubject

inherited

Definition at line 127 of file SGObject.h.

◆ SGSubscriber

typedef rxcpp::subscriber< ObservedValue, rxcpp::observer<ObservedValue, void, void, void, void> > SGSubscriber

inherited

Definition at line 133 of file SGObject.h.

Constructor & Destructor Documentation

◆ CMMD() [1/2]

CMMD ( )

Default constructor

Definition at line 66 of file MMD.cpp.

◆ CMMD() [2/2]

CMMD	(	CFeatures *	samples_from_p,
		CFeatures *	samples_from_q
	)

Convenience constructor that initializes the samples from two distributions.

Parameters

samples_from_p	Samples from \(p\)
samples_from_q	Samples from \(q\)

Definition at line 71 of file MMD.cpp.

◆ ~CMMD()

~CMMD ( )

virtual

Destructor

Definition at line 84 of file MMD.cpp.

Member Function Documentation

◆ add_kernel()

void add_kernel ( CKernel * kernel )

Method that adds a kernel instance to be used for kernel selection. Please note that the kernels added by this method are NOT set as the main test kernel unless select_kernel() method is executed.

This method is NOT thread safe. Please DO NOT use this method from multiple threads.

Parameters

kernel One of the kernel instances with which learning algorithm will work.

Definition at line 109 of file MMD.cpp.

◆ build_gradient_parameter_dictionary()

void build_gradient_parameter_dictionary ( CMap< TParameter *, CSGObject *> * dict )

inherited

Builds a dictionary of all parameters in SGObject as well of those of SGObjects that are parameters of this object. Dictionary maps parameters to the objects that own them.

Parameters

dict	dictionary of parameters to be built.

Definition at line 635 of file SGObject.cpp.

◆ cleanup()

void cleanup ( )

Method that releases the pre-computed kernel that is used in the computation.

Definition at line 124 of file MMD.cpp.

◆ clone()

CSGObject * clone ( )

virtualinherited

Creates a clone of the current object. This is done via recursively traversing all parameters, which corresponds to a deep copy. Calling equals on the cloned object always returns true although none of the memory of both objects overlaps.

Returns: an identical copy of the given object, which is disjoint in memory. NULL if the clone fails. Note that the returned object is SG_REF'ed

Reimplemented in CDynamicArray< T >, CDynamicArray< float64_t >, CDynamicArray< float32_t >, CDynamicArray< int32_t >, CDynamicArray< char >, CDynamicArray< bool >, CDynamicObjectArray, CAlphabet, and CMKL.

Definition at line 734 of file SGObject.cpp.

◆ clone_parameters()

bool clone_parameters ( CSGObject * other )

protectedinherited

Definition at line 759 of file SGObject.cpp.

◆ compute_distance()

CCustomDistance * compute_distance ( CDistance * distance )

inherited

Method that pre-computes the pair-wise distance between the samples using the provided distance instance.

Parameters

distance The distance instance used for pre-computing the pair-wise distance.

Returns: A newly created CCustomDistance instance representing the pre-computed pair-wise distance between the samples.

Definition at line 99 of file TwoDistributionTest.cpp.

◆ compute_joint_distance()

CCustomDistance * compute_joint_distance ( CDistance * distance )

inherited

Method that pre-computes the pair-wise distance between the joint samples using the provided distance instance. A temporary object appending the samples from both the distributions is created in order to perform the task.

Parameters

distance The distance instance used for pre-computing the pair-wise distance.

Returns: A newly created CCustomDistance instance representing the pre-computed pair-wise distance between the joint samples.

Definition at line 128 of file TwoDistributionTest.cpp.

◆ compute_p_value()

float64_t compute_p_value ( float64_t statistic )

virtualinherited

Method that computes a p-value based on current method for approximating the null-distribution. The p-value is the 1-p quantile of the null- distribution where the given statistic lies in.

This method depends on the implementation of sample_null method which should be implemented by the sub-classes.

Parameters

statistic statistic value to compute the p-value for

Returns: p-value parameter statistic is the (1-p) percentile of the null distribution

Reimplemented in CQuadraticTimeMMD, CBTestMMD, and CLinearTimeMMD.

Definition at line 77 of file HypothesisTest.cpp.

◆ compute_statistic()

virtual float64_t compute_statistic ( )

pure virtual

Interface for computing the test-statistic for the hypothesis test.

Returns: test statistic for the given data/parameters/methods

Implements CTwoSampleTest.

Implemented in CQuadraticTimeMMD, and CStreamingMMD.

◆ compute_threshold()

float64_t compute_threshold ( float64_t alpha )

virtualinherited

Method that computes a threshold based on current method for approximating the null-distribution. The threshold is the value that a statistic has to have in ordner to reject the null-hypothesis.

This method depends on the implementation of sample_null method which should be implemented by the sub-classes.

Parameters

alpha test level to reject null-hypothesis

Returns: Threshold for statistics to reject null-hypothesis

Reimplemented in CQuadraticTimeMMD, CBTestMMD, and CLinearTimeMMD.

Definition at line 85 of file HypothesisTest.cpp.

◆ deep_copy()

CSGObject * deep_copy ( ) const

virtualinherited

A deep copy. All the instance variables will also be copied.

Definition at line 232 of file SGObject.cpp.

◆ equals()

bool equals	(	CSGObject *	other,
		float64_t	accuracy = `0.0`,
		bool	tolerant = `false`
	)

virtualinherited

Recursively compares the current SGObject to another one. Compares all registered numerical parameters, recursion upon complex (SGObject) parameters. Does not compare pointers!

May be overwritten but please do with care! Should not be necessary in most cases.

Parameters

other	object to compare with
accuracy	accuracy to use for comparison (optional)
tolerant	allows linient check on float equality (within accuracy)

Returns: true if all parameters were equal, false if not

Definition at line 656 of file SGObject.cpp.

◆ get() [1/2]

T get ( const Tag< T > & _tag ) const

inherited

Getter for a class parameter, identified by a Tag. Throws an exception if the class does not have such a parameter.

Parameters

_tag	name and type information of parameter

Returns: value of the parameter identified by the input tag

Definition at line 381 of file SGObject.h.

◆ get() [2/2]

T get ( const std::string & name ) const

inherited

Getter for a class parameter, identified by a name. Throws an exception if the class does not have such a parameter.

Parameters

name	name of the parameter

Returns: value of the parameter corresponding to the input name and type

Definition at line 404 of file SGObject.h.

◆ get_data_mgr() [1/2]

DataManager & get_data_mgr ( )

protectedinherited

Definition at line 104 of file HypothesisTest.cpp.

◆ get_data_mgr() [2/2]

const DataManager & get_data_mgr ( ) const

protectedinherited

Definition at line 109 of file HypothesisTest.cpp.

◆ get_global_io()

SGIO * get_global_io ( )

inherited

get the io object

Returns: io object

Definition at line 269 of file SGObject.cpp.

◆ get_global_parallel()

Parallel * get_global_parallel ( )

inherited

get the parallel object

Returns: parallel object

Definition at line 311 of file SGObject.cpp.

◆ get_global_version()

Version * get_global_version ( )

inherited

get the version object

Returns: version object

Definition at line 324 of file SGObject.cpp.

◆ get_kernel()

CKernel * get_kernel ( ) const

inherited

Returns: The kernel instance that is presently being used for performing the test

Definition at line 73 of file TwoSampleTest.cpp.

◆ get_kernel_mgr() [1/2]

KernelManager & get_kernel_mgr ( )

protectedinherited

Definition at line 83 of file TwoSampleTest.cpp.

◆ get_kernel_mgr() [2/2]

const KernelManager & get_kernel_mgr ( ) const

protectedinherited

Definition at line 88 of file TwoSampleTest.cpp.

◆ get_kernel_selection_strategy()

CKernelSelectionStrategy const * get_kernel_selection_strategy ( ) const

Method that returns the kernel selection strategy wrapper object that will be/ was used in the last kernel learning algorithm. Use this method when results of intermediate steps taken by the kernel selection algorithms are of interest.

Returns: The internal instance of CKernelSelectionStrategy that holds intermediate measures computed at the time of the last kernel selection algorithm invocation.

Definition at line 104 of file MMD.cpp.

◆ get_modelsel_names()

SGStringList< char > get_modelsel_names ( )

inherited

Returns: vector of names of all parameters which are registered for model selection

Definition at line 536 of file SGObject.cpp.

◆ get_modsel_param_descr()

char * get_modsel_param_descr ( const char * param_name )

inherited

Returns description of a given parameter string, if it exists. SG_ERROR otherwise

Parameters

param_name name of the parameter

Returns: description of the parameter

Definition at line 560 of file SGObject.cpp.

◆ get_modsel_param_index()

index_t get_modsel_param_index ( const char * param_name )

inherited

Returns index of model selection parameter with provided index

Parameters

param_name name of model selection parameter

Returns: index of model selection parameter with provided name, -1 if there is no such

Definition at line 573 of file SGObject.cpp.

◆ get_name()

const char * get_name ( ) const

virtual

Returns: The name of this class

Reimplemented from CTwoSampleTest.

Reimplemented in CQuadraticTimeMMD, CStreamingMMD, CBTestMMD, and CLinearTimeMMD.

Definition at line 159 of file MMD.cpp.

◆ get_null_approximation_method()

ENullApproximationMethod get_null_approximation_method ( ) const

Returns: The null-approximation method

Definition at line 154 of file MMD.cpp.

◆ get_num_null_samples()

index_t get_num_null_samples ( ) const

Returns: Number of null-samples

Definition at line 134 of file MMD.cpp.

◆ get_num_samples_p()

const index_t get_num_samples_p ( ) const

inherited

Returns: The number of samples from \(\mathbf{P}\).

Definition at line 81 of file TwoDistributionTest.cpp.

◆ get_num_samples_q()

const index_t get_num_samples_q ( ) const

inherited

Returns: The number of samples from \(\mathbf{Q}\).

Definition at line 93 of file TwoDistributionTest.cpp.

◆ get_p()

CFeatures * get_p ( ) const

inherited

Returns: The samples from \(\mathbf{P}\).

Definition at line 56 of file TwoDistributionTest.cpp.

◆ get_parameters_observable()

SGObservable* get_parameters_observable ( )

inherited

Get parameters observable

Returns: RxCpp observable

Definition at line 415 of file SGObject.h.

◆ get_q()

CFeatures * get_q ( ) const

inherited

Returns: The samples from \(\mathbf{Q}\).

Definition at line 69 of file TwoDistributionTest.cpp.

◆ get_statistic_type()

EStatisticType get_statistic_type ( ) const

Returns: The type of the estimator for MMD^2

Definition at line 144 of file MMD.cpp.

◆ has() [1/3]

bool has ( const std::string & name ) const

inherited

Checks if object has a class parameter identified by a name.

Parameters

name	name of the parameter

Returns: true if the parameter exists with the input name

Definition at line 304 of file SGObject.h.

◆ has() [2/3]

bool has ( const Tag< T > & tag ) const

inherited

Checks if object has a class parameter identified by a Tag.

Parameters

tag	tag of the parameter containing name and type information

Returns: true if the parameter exists with the input tag

Definition at line 315 of file SGObject.h.

◆ has() [3/3]

bool has ( const std::string & name ) const

inherited

Checks if a type exists for a class parameter identified by a name.

Parameters

name	name of the parameter

Returns: true if the parameter exists with the input name and type

Definition at line 326 of file SGObject.h.

◆ is_generic()

bool is_generic ( EPrimitiveType * generic ) const

virtualinherited

If the SGSerializable is a class template then TRUE will be returned and GENERIC is set to the type of the generic.

Parameters

generic set to the type of the generic if returning TRUE

Returns: TRUE if a class template.

Definition at line 330 of file SGObject.cpp.

◆ list_observable_parameters()

void list_observable_parameters ( )

inherited

Print to stdout a list of observable parameters

Definition at line 878 of file SGObject.cpp.

◆ load_serializable()

bool load_serializable	(	CSerializableFile *	file,
		const char *	prefix = `""`
	)

virtualinherited

Load this object from file. If it will fail (returning FALSE) then this object will contain inconsistent data and should not be used!

Parameters

file	where to load from
prefix	prefix for members

Returns: TRUE if done, otherwise FALSE

Definition at line 403 of file SGObject.cpp.

◆ load_serializable_post()

void load_serializable_post	(		)
throw	(	ShogunException
	)

protectedvirtualinherited

Can (optionally) be overridden to post-initialize some member variables which are not PARAMETER::ADD'ed. Make sure that at first the overridden method BASE_CLASS::LOAD_SERIALIZABLE_POST is called.

Exceptions

ShogunException will be thrown if an error occurs.

Reimplemented in CKernel, CWeightedDegreePositionStringKernel, CList, CAlphabet, CLinearHMM, CGaussianKernel, CInverseMultiQuadricKernel, CCircularKernel, and CExponentialKernel.

Definition at line 460 of file SGObject.cpp.

◆ load_serializable_pre()

void load_serializable_pre	(		)
throw	(	ShogunException
	)

protectedvirtualinherited

Can (optionally) be overridden to pre-initialize some member variables which are not PARAMETER::ADD'ed. Make sure that at first the overridden method BASE_CLASS::LOAD_SERIALIZABLE_PRE is called.

Exceptions

ShogunException will be thrown if an error occurs.

Reimplemented in CDynamicArray< T >, CDynamicArray< float64_t >, CDynamicArray< float32_t >, CDynamicArray< int32_t >, CDynamicArray< char >, CDynamicArray< bool >, and CDynamicObjectArray.

Definition at line 455 of file SGObject.cpp.

◆ normalize_statistic()

virtual float64_t normalize_statistic ( float64_t statistic ) const

protectedpure virtual

Implemented in CQuadraticTimeMMD, and CStreamingMMD.

◆ observe()

void observe ( const ObservedValue value )

protectedinherited

Observe a parameter value and emit them to observer.

Parameters

value Observed parameter's value

Definition at line 828 of file SGObject.cpp.

◆ parameter_hash_changed()

bool parameter_hash_changed ( )

virtualinherited

Returns: whether parameter combination has changed since last update

Definition at line 296 of file SGObject.cpp.

◆ perform_test()

bool perform_test ( float64_t alpha )

inherited

Method that performs the complete hypothesis test on current data and returns a binary answer: wheter null hypothesis is rejected or not.

This is just a wrapper for the above compute_p_value() method that returns a p-value. If this p-value lies below the test level alpha, the null hypothesis is rejected.

Should not be overwritten in subclasses. (Therefore not virtual)

Parameters

alpha test level alpha.

Returns: true if null hypothesis is rejected and false otherwise

Definition at line 92 of file HypothesisTest.cpp.

◆ print_modsel_params()

void print_modsel_params ( )

inherited

prints all parameter registered for model selection and their type

Definition at line 512 of file SGObject.cpp.

◆ print_serializable()

void print_serializable ( const char * prefix = "" )

virtualinherited

prints registered parameters out

Parameters

prefix prefix for members

Definition at line 342 of file SGObject.cpp.

◆ ref()

int32_t ref ( )

inherited

increase reference counter

Returns: reference count

Definition at line 186 of file SGObject.cpp.

◆ ref_count()

int32_t ref_count ( )

inherited

display reference counter

Returns: reference count

Definition at line 193 of file SGObject.cpp.

◆ register_observable_param()

void register_observable_param	(	const std::string &	name,
		const SG_OBS_VALUE_TYPE	type,
		const std::string &	description
	)

protectedinherited

Register which params this object can emit.

Parameters

name	the param name
type	the param type
description	a user oriented description

Definition at line 871 of file SGObject.cpp.

◆ register_param() [1/2]

void register_param	(	Tag< T > &	_tag,
		const T &	value
	)

protectedinherited

Registers a class parameter which is identified by a tag. This enables the parameter to be modified by set() and retrieved by get(). Parameters can be registered in the constructor of the class.

Parameters

_tag	name and type information of parameter
value	value of the parameter

Definition at line 472 of file SGObject.h.

◆ register_param() [2/2]

void register_param	(	const std::string &	name,
		const T &	value
	)

protectedinherited

Registers a class parameter which is identified by a name. This enables the parameter to be modified by set() and retrieved by get(). Parameters can be registered in the constructor of the class.

Parameters

name	name of the parameter
value	value of the parameter along with type information

Definition at line 485 of file SGObject.h.

◆ sample_null()

virtual SGVector<float64_t> sample_null ( )

pure virtual

Interface for computing the samples under the null-hypothesis.

Returns: vector of all statistics

Implements CTwoSampleTest.

Implemented in CQuadraticTimeMMD, and CStreamingMMD.

◆ save_serializable()

bool save_serializable	(	CSerializableFile *	file,
		const char *	prefix = `""`
	)

virtualinherited

Save this object to file.

Parameters

file	where to save the object; will be closed during returning if PREFIX is an empty string.
prefix	prefix for members

Returns: TRUE if done, otherwise FALSE

Definition at line 348 of file SGObject.cpp.

◆ save_serializable_post()

void save_serializable_post	(		)
throw	(	ShogunException
	)

protectedvirtualinherited

Can (optionally) be overridden to post-initialize some member variables which are not PARAMETER::ADD'ed. Make sure that at first the overridden method BASE_CLASS::SAVE_SERIALIZABLE_POST is called.

Exceptions

ShogunException will be thrown if an error occurs.

Reimplemented in CKernel.

Definition at line 470 of file SGObject.cpp.

◆ save_serializable_pre()

void save_serializable_pre	(		)
throw	(	ShogunException
	)

protectedvirtualinherited

Can (optionally) be overridden to pre-initialize some member variables which are not PARAMETER::ADD'ed. Make sure that at first the overridden method BASE_CLASS::SAVE_SERIALIZABLE_PRE is called.

Exceptions

ShogunException will be thrown if an error occurs.

Reimplemented in CKernel, CDynamicArray< T >, CDynamicArray< float64_t >, CDynamicArray< float32_t >, CDynamicArray< int32_t >, CDynamicArray< char >, CDynamicArray< bool >, and CDynamicObjectArray.

Definition at line 465 of file SGObject.cpp.

◆ select_kernel()

void select_kernel ( )

virtual

Method that selects/learns the kernel based on the defined kernel selection strategy. If no explicit kernel selection strategy was set using set_kernel_selection_strategy() method, then a default strategy is used. Please see EKernelSelectionMethod for the default strategy.

This method is NOT thread safe. It replaces the internel kernel set by set_kernel() method, if there was any. Please DO NOT use this method from multiple threads.

The learned/selected kernel can be obtained from a subsequent get_kernel() call.

This method expects train-test mode to be turned on at the time of invocation. Please see the class documentation of CHypothesisTest.

Reimplemented in CQuadraticTimeMMD.

Definition at line 114 of file MMD.cpp.

◆ set() [1/2]

void set	(	const Tag< T > &	_tag,
		const T &	value
	)

inherited

Setter for a class parameter, identified by a Tag. Throws an exception if the class does not have such a parameter.

Parameters

_tag	name and type information of parameter
value	value of the parameter

Definition at line 342 of file SGObject.h.

◆ set() [2/2]

void set	(	const std::string &	name,
		const T &	value
	)

inherited

Setter for a class parameter, identified by a name. Throws an exception if the class does not have such a parameter.

Parameters

name	name of the parameter
value	value of the parameter along with type information

Definition at line 368 of file SGObject.h.

◆ set_generic() [1/16]

void set_generic ( )

inherited

Definition at line 73 of file SGObject.cpp.

◆ set_generic() [2/16]

void set_generic ( )

inherited

Definition at line 78 of file SGObject.cpp.

◆ set_generic() [3/16]

void set_generic ( )

inherited

Definition at line 83 of file SGObject.cpp.

◆ set_generic() [4/16]

void set_generic ( )

inherited

Definition at line 88 of file SGObject.cpp.

◆ set_generic() [5/16]

void set_generic ( )

inherited

Definition at line 93 of file SGObject.cpp.

◆ set_generic() [6/16]

void set_generic ( )

inherited

Definition at line 98 of file SGObject.cpp.

◆ set_generic() [7/16]

void set_generic ( )

inherited

Definition at line 103 of file SGObject.cpp.

◆ set_generic() [8/16]

void set_generic ( )

inherited

Definition at line 108 of file SGObject.cpp.

◆ set_generic() [9/16]

void set_generic ( )

inherited

Definition at line 113 of file SGObject.cpp.

◆ set_generic() [10/16]

void set_generic ( )

inherited

Definition at line 118 of file SGObject.cpp.

◆ set_generic() [11/16]

void set_generic ( )

inherited

Definition at line 123 of file SGObject.cpp.

◆ set_generic() [12/16]

void set_generic ( )

inherited

Definition at line 128 of file SGObject.cpp.

◆ set_generic() [13/16]

void set_generic ( )

inherited

Definition at line 133 of file SGObject.cpp.

◆ set_generic() [14/16]

void set_generic ( )

inherited

Definition at line 138 of file SGObject.cpp.

◆ set_generic() [15/16]

void set_generic ( )

inherited

Definition at line 143 of file SGObject.cpp.

◆ set_generic() [16/16]

void set_generic ( )

inherited

set generic type to T

◆ set_global_io()

void set_global_io ( SGIO * io )

inherited

set the io object

Parameters

io	io object to use

Definition at line 262 of file SGObject.cpp.

◆ set_global_parallel()

void set_global_parallel ( Parallel * parallel )

inherited

set the parallel object

Parameters

parallel parallel object to use

Definition at line 275 of file SGObject.cpp.

◆ set_global_version()

void set_global_version ( Version * version )

inherited

set the version object

Parameters

version version object to use

Definition at line 317 of file SGObject.cpp.

◆ set_kernel()

void set_kernel ( CKernel * kernel )

virtualinherited

Method that sets the kernel that is used for performing the two-sample test. It is kept virtual so that sub-classes can perform other initialization tasks that has to be trigger every time a kernel is set/updated.

Parameters

kernel The kernel instance.

Reimplemented in CQuadraticTimeMMD.

Definition at line 66 of file TwoSampleTest.cpp.

◆ set_kernel_selection_strategy() [1/2]

void set_kernel_selection_strategy	(	EKernelSelectionMethod	method,
		bool	weighted = `false`
	)

Method that sets the specific kernel selection strategy based on the specific parameters provided. Please see class documentation for details. Use this method for every other strategy other than KSM_CROSS_VALIDATION.

Parameters

method	The kernel selection method as specified in EKernelSelectionMethod.
weighted	If true, then an weighted combination of the kernel is used after solving an optimization. If false, only a single kernel is selected among the provided ones.

Definition at line 89 of file MMD.cpp.

◆ set_kernel_selection_strategy() [2/2]

void set_kernel_selection_strategy	(	EKernelSelectionMethod	method,
		index_t	num_runs,
		index_t	num_folds,
		float64_t	alpha
	)

Method that sets the specific kernel selection strategy based on the specific parameters provided. Please see class documentation for details. Use this method for KSM_CROSS_VALIDATION.

Parameters

method	The kernel selection method as specified in EKernelSelectionMethod.
num_runs	The number of total runs of the cross-validation algorithm.
num_folds	The number of folds (k) to be used in k-fold stratified cross-validation.
alpha	The threshold to be used while performing test for the test-folds.

Definition at line 95 of file MMD.cpp.

◆ set_null_approximation_method()

void set_null_approximation_method ( ENullApproximationMethod nmethod )

Method that sets the approach to be taken while approximating the null-samples.

The null-approximation method

Definition at line 149 of file MMD.cpp.

◆ set_num_null_samples()

void set_num_null_samples ( index_t null_samples )

Method that sets the number of null-samples used for computing p-value.

Parameters

null_samples Number of null-samples.

Definition at line 129 of file MMD.cpp.

◆ set_num_samples_p()

void set_num_samples_p ( index_t num_samples_from_p )

inherited

Method that initializes the number of samples to be drawn from distribution \(\mathbf{P}\). Please ensure to call this method if you are intending to use streaming data generators that generate the samples on the fly. For other types of features, the number of samples is set internally from the features object itself, therefore this method should not be used.

Parameters

num_samples_from_p The CFeatures instance representing the samples from \(\mathbf{P}\).

Definition at line 75 of file TwoDistributionTest.cpp.

◆ set_num_samples_q()

void set_num_samples_q ( index_t num_samples_from_q )

inherited

Method that initializes the number of samples to be drawn from distribution \(\mathbf{Q}\). Please ensure to call this method if you are intending to use streaming data generators that generate the samples on the fly. For other types of features, the number of samples is set internally from the features object itself, therefore this method should not be used.

Parameters

num_samples_from_q The CFeatures instance representing the samples from \(\mathbf{Q}\).

Definition at line 87 of file TwoDistributionTest.cpp.

◆ set_p()

void set_p ( CFeatures * samples_from_p )

virtualinherited

Method that initializes the samples from \(\mathbf{P}\). This method is kept virtual for the sub-classes to perform additional initialization tasks that have to be performed every time features are set/updated.

Parameters

samples_from_p The CFeatures instance representing the samples from \(\mathbf{P}\).

Reimplemented in CQuadraticTimeMMD.

Definition at line 49 of file TwoDistributionTest.cpp.

◆ set_q()

void set_q ( CFeatures * samples_from_q )

virtualinherited

Method that initializes the samples from \(\mathbf{Q}\). This method is kept virtual for the sub-classes to perform additional initialization tasks that have to be performed every time features are set/updated.

Parameters

samples_from_q The CFeatures instance representing the samples from \(\mathbf{Q}\).

Reimplemented in CQuadraticTimeMMD.

Definition at line 62 of file TwoDistributionTest.cpp.

◆ set_statistic_type()

void set_statistic_type ( EStatisticType stype )

Method that sets the type of the estimator for MMD^2

Parameters

stype The type of the estimator for MMD^2

Definition at line 139 of file MMD.cpp.

◆ set_train_test_mode()

void set_train_test_mode ( bool on )

inherited

Method that enables/disables the training-testing mode. If this option is turned on, then the samples would be split in two pieces: one chunk would be used for training algorithms and the other chunk would be used for performing tests. If this option is turned off, the entire data would be used for performing the test. Before running any training algorithms, make sure to turn this mode on.

By default, the training-testing mode is turned off.

See also: {set_train_test_ratio()}

Parameters

on	Whether to enable/disable the training-testing mode

Definition at line 66 of file HypothesisTest.cpp.

◆ set_train_test_ratio()

void set_train_test_ratio ( float64_t ratio )

inherited

Method that specifies the ratio of training-testing data split for the algorithms. Note that this is NOT the percentage of samples to be used for training, rather the ratio of the number of samples to be used for training and that of testing.

By default, an equal 50-50 split (ratio = 1) is made.

See also: {set_train_test_mode()}

Parameters

ratio The ratio of the number of samples to be used for training and that of testing

Definition at line 71 of file HypothesisTest.cpp.

◆ shallow_copy()

CSGObject * shallow_copy ( ) const

virtualinherited

A shallow copy. All the SGObject instance variables will be simply assigned and SG_REF-ed.

Reimplemented in CGaussianKernel.

Definition at line 226 of file SGObject.cpp.

◆ subscribe_to_parameters()

void subscribe_to_parameters ( ParameterObserverInterface * obs )

inherited

Subscribe a parameter observer to watch over params

Definition at line 811 of file SGObject.cpp.

◆ unref()

int32_t unref ( )

inherited

decrement reference counter and deallocate object if refcount is zero before or after decrementing it

Returns: reference count

Definition at line 200 of file SGObject.cpp.

◆ unset_generic()

void unset_generic ( )

inherited

unset generic type

this has to be called in classes specializing a template class

Definition at line 337 of file SGObject.cpp.

◆ update_parameter_hash()

void update_parameter_hash ( )

virtualinherited

Updates the hash of current parameter combination

Definition at line 282 of file SGObject.cpp.

Member Data Documentation

◆ io

SGIO* io

inherited

io

Definition at line 600 of file SGObject.h.

◆ m_gradient_parameters

Parameter* m_gradient_parameters

inherited

parameters wrt which we can compute gradients

Definition at line 615 of file SGObject.h.

◆ m_hash

uint32_t m_hash

inherited

Hash of parameter values

Definition at line 618 of file SGObject.h.

◆ m_model_selection_parameters

Parameter* m_model_selection_parameters

inherited

model selection parameters

Definition at line 612 of file SGObject.h.

◆ m_parameters

Parameter* m_parameters

inherited

parameters

Definition at line 609 of file SGObject.h.

◆ parallel

Parallel* parallel

inherited

parallel

Definition at line 603 of file SGObject.h.

◆ version

Version* version

inherited

version

Definition at line 606 of file SGObject.h.

The documentation for this class was generated from the following files:

Detailed Description

Classes

Public Types

Public Member Functions

Public Attributes

Protected Member Functions

Member Typedef Documentation

◆ SGObservable

◆ SGSubject

◆ SGSubscriber

Constructor & Destructor Documentation

◆ CMMD() [1/2]

◆ CMMD() [2/2]

◆ ~CMMD()

Member Function Documentation

◆ add_kernel()

◆ build_gradient_parameter_dictionary()

◆ cleanup()

◆ clone()

◆ clone_parameters()

◆ compute_distance()

◆ compute_joint_distance()

◆ compute_p_value()

◆ compute_statistic()

◆ compute_threshold()

◆ deep_copy()

◆ equals()

◆ get() [1/2]

◆ get() [2/2]

◆ get_data_mgr() [1/2]

◆ get_data_mgr() [2/2]

◆ get_global_io()

◆ get_global_parallel()

◆ get_global_version()

◆ get_kernel()

◆ get_kernel_mgr() [1/2]

◆ get_kernel_mgr() [2/2]

◆ get_kernel_selection_strategy()

◆ get_modelsel_names()

◆ get_modsel_param_descr()

◆ get_modsel_param_index()

◆ get_name()

◆ get_null_approximation_method()

◆ get_num_null_samples()

◆ get_num_samples_p()

◆ get_num_samples_q()

◆ get_p()

◆ get_parameters_observable()

◆ get_q()

◆ get_statistic_type()

◆ has() [1/3]

◆ has() [2/3]

◆ has() [3/3]

◆ is_generic()

◆ list_observable_parameters()

◆ load_serializable()

◆ load_serializable_post()

◆ load_serializable_pre()

◆ normalize_statistic()

◆ observe()

◆ parameter_hash_changed()

◆ perform_test()

◆ print_modsel_params()

◆ print_serializable()

◆ ref()

◆ ref_count()

◆ register_observable_param()

◆ register_param() [1/2]

◆ register_param() [2/2]

◆ sample_null()

◆ save_serializable()

◆ save_serializable_post()

◆ save_serializable_pre()

◆ select_kernel()

◆ set() [1/2]

◆ set() [2/2]

◆ set_generic() [1/16]

◆ set_generic() [2/16]

◆ set_generic() [3/16]

◆ set_generic() [4/16]