The base class for all the independence tests used for learning. More...

#include <agrum/BN/learning/scores/independenceTest.h>

Inheritance diagram for gum::learning::IndependenceTest:

Collaboration diagram for gum::learning::IndependenceTest:

Public Member Functions
Constructors / Destructors
	IndependenceTest (const DBRowGeneratorParser &parser, const Prior &external_prior, const std::vector< std::pair< std::size_t, std::size_t > > &ranges, const Bijection< NodeId, std::size_t > &nodeId2columns=Bijection< NodeId, std::size_t >())
	default constructor
	IndependenceTest (const DBRowGeneratorParser &parser, const Prior &external_prior, const Bijection< NodeId, std::size_t > &nodeId2columns=Bijection< NodeId, std::size_t >())
	default constructor
virtual IndependenceTest *	clone () const =0
	virtual copy constructor
	~IndependenceTest () override
	destructor
Statistics
virtual std::pair< double, double >	statistics (NodeId var1, NodeId var2, const std::vector< NodeId > &rhs_ids={})=0
	returns the pair (test statistic, p-value) for the independence test X _\|_ Y \| Z
Accessors / Modifiers
void	setNumberOfThreads (Size nb) override
	sets the number max of threads that can be used
Size	getNumberOfThreads () const override
	returns the current max number of threads of the scheduler
bool	isGumNumberOfThreadsOverriden () const override
	indicates whether the user set herself the number of threads
virtual void	setMinNbRowsPerThread (const std::size_t nb) const
	changes the number min of rows a thread should process in a multithreading context
virtual std::size_t	minNbRowsPerThread () const
	returns the minimum of rows that each thread should process
void	setRanges (const std::vector< std::pair< std::size_t, std::size_t > > &new_ranges)
	sets new ranges to perform the counts
void	clearRanges ()
	reset the ranges to the one range corresponding to the whole database
const std::vector< std::pair< std::size_t, std::size_t > > &	ranges () const
	returns the current ranges
virtual void	clear ()
	clears all the data structures from memory, including the cache
virtual void	clearCache ()
	clears the current cache
virtual void	useCache (const bool on_off)
	turn on/off the use of a cache of the previously computed score
const Bijection< NodeId, std::size_t > &	nodeId2Columns () const
	return the mapping between the columns of the database and the node ids
const DatabaseTable &	database () const
	return the database used by the score

Protected Member Functions
	IndependenceTest (const IndependenceTest &from)
	copy constructor
	IndependenceTest (IndependenceTest &&from)
	move constructor
IndependenceTest &	operator= (const IndependenceTest &from)
	copy operator
IndependenceTest &	operator= (IndependenceTest &&from)
	move operator
std::vector< double >	marginalize_ (const std::size_t node_2_marginalize, const std::size_t X_size, const std::size_t Y_size, const std::size_t Z_size, const std::vector< double > &N_xyz) const
	returns a counting vector where variables are marginalized from N_xyz
template<typename CellContribFn>
std::pair< double, double >	computeStatistics_ (const IdCondSet &idset, CellContribFn cellContrib)
	shared loop for chi-squared-family statistics

Static Protected Member Functions
static Size	degreesOfFreedom_ (std::size_t X_size, std::size_t Y_size, std::size_t Z_size=1, std::size_t n_skipped=0)
	returns the degrees of freedom for a chi2/G2 test X _\|_ Y \| Z

Protected Attributes
std::vector< std::size_t >	_domain_sizes_
	the domain sizes of the variables (indexed by column id in the database)
const double	one_log2_ {M_LOG2E}
	1 / log(2)
Prior *	prior_ {nullptr}
	the expert knowledge prior added to the contingency tables
RecordCounter	counter_
	the record counter used for the counts over discrete variables
ScoringCache	cache_
	the scoring cache
bool	use_cache_ {true}
	a Boolean indicating whether we wish to use the cache
const std::vector< NodeId >	empty_ids_
	an empty vector

Detailed Description

The base class for all the independence tests used for learning.

Definition at line 61 of file independenceTest.h.

Constructor & Destructor Documentation

◆ IndependenceTest() [1/4]

gum::learning::IndependenceTest::IndependenceTest	(	const DBRowGeneratorParser &	parser,
		const Prior &	external_prior,
		const std::vector< std::pair< std::size_t, std::size_t > > &	ranges,
		const Bijection< NodeId, std::size_t > &	nodeId2columns = Bijection< NodeId, std::size_t >() )

default constructor

Parameters

parser	the parser used to parse the database
external_prior	An prior that we add to the computation of the score (this should come from expert knowledge): this consists in adding numbers to counts in the contingency tables
ranges	a set of pairs {(X1,Y1),...,(Xn,Yn)} of database's rows indices. The counts are then performed only on the union of the rows [Xi,Yi), i in {1,...,n}. This is useful, e.g, when performing cross validation tasks, in which part of the database should be ignored. An empty set of ranges is equivalent to an interval [X,Y) ranging over the whole database.
nodeId2Columns	a mapping from the ids of the nodes in the graphical model to the corresponding column in the DatabaseTable parsed by the parser. This enables estimating from a database in which variable A corresponds to the 2nd column the parameters of a BN in which variable A has a NodeId of 5. An empty nodeId2Columns bijection means that the mapping is an identity, i.e., the value of a NodeId is equal to the index of the column in the DatabaseTable.

Warning: If nodeId2columns is not empty, then only the statistics over the ids belonging to this bijection can be computed: applying method statistics() over other ids will raise exception NotFound.

References gum::learning::CachedContingencyCounter::ranges().

Referenced by IndependenceTest(), IndependenceTest(), clone(), operator=(), and operator=().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ IndependenceTest() [2/4]

gum::learning::IndependenceTest::IndependenceTest	(	const DBRowGeneratorParser &	parser,
		const Prior &	external_prior,
		const Bijection< NodeId, std::size_t > &	nodeId2columns = Bijection< NodeId, std::size_t >() )

default constructor

Parameters

parser	the parser used to parse the database
external_prior	An prior that we add to the computation of the score (this should come from expert knowledge): this consists in adding numbers to counts in the contingency tables
nodeId2Columns	a mapping from the ids of the nodes in the graphical model to the corresponding column in the DatabaseTable parsed by the parser. This enables estimating from a database in which variable A corresponds to the 2nd column the parameters of a BN in which variable A has a NodeId of 5. An empty nodeId2Columns bijection means that the mapping is an identity, i.e., the value of a NodeId is equal to the index of the column in the DatabaseTable.

Warning: If nodeId2columns is not empty, then only the statistics over the ids belonging to this bijection can be computed: applying method statistics() over other ids will raise exception NotFound.

◆ ~IndependenceTest()

gum::learning::IndependenceTest::~IndependenceTest ( )

override

destructor

◆ IndependenceTest() [3/4]

gum::learning::IndependenceTest::IndependenceTest ( const IndependenceTest & from )

protected

copy constructor

References IndependenceTest().

Here is the call graph for this function:

◆ IndependenceTest() [4/4]

gum::learning::IndependenceTest::IndependenceTest ( IndependenceTest && from )

protected

move constructor

References IndependenceTest().

Here is the call graph for this function:

Member Function Documentation

◆ clear()

virtual void gum::learning::CachedContingencyCounter::clear ( )

virtualinherited

clears all the data structures from memory, including the cache

Reimplemented in gum::learning::KNML.

◆ clearCache()

virtual void gum::learning::CachedContingencyCounter::clearCache ( )

virtualinherited

clears the current cache

Reimplemented in gum::learning::KNML.

◆ clearRanges()

void gum::learning::CachedContingencyCounter::clearRanges ( )

inherited

reset the ranges to the one range corresponding to the whole database

◆ clone()

virtual IndependenceTest * gum::learning::IndependenceTest::clone ( ) const

nodiscardpure virtual

virtual copy constructor

Implemented in gum::learning::IndepTestChi2, and gum::learning::IndepTestG2.

References IndependenceTest().

Here is the call graph for this function:

◆ computeStatistics_()

template<typename CellContribFn>

std::pair< double, double > gum::learning::IndependenceTest::computeStatistics_	(	const IdCondSet &	idset,
		CellContribFn	cellContrib )

protected

shared loop for chi-squared-family statistics

Handles counts, marginalisation, structural-zero detection and the final p-value. Only active cells (margX * margY != 0) are forwarded to cellContrib; sampling zeros must be handled inside the lambda. Signature: double cellContrib(double O, double margX, double margY, double total)

◆ database()

const DatabaseTable & gum::learning::CachedContingencyCounter::database ( ) const

inherited

return the database used by the score

◆ degreesOfFreedom_()

Size gum::learning::IndependenceTest::degreesOfFreedom_	(	std::size_t	X_size,
		std::size_t	Y_size,
		std::size_t	Z_size = 1,
		std::size_t	n_skipped = 0 )

staticprotected

returns the degrees of freedom for a chi2/G2 test X _|_ Y | Z

Parameters

X_size	domain size of X
Y_size	domain size of Y
Z_size	product of domain sizes of the conditioning variables (1 if no conditioning set)
n_skipped	number of cells excluded from the statistic sum because their expected count is zero (silent cells). Each such cell effectively removes one degree of freedom. The result is clamped to 1 to keep the distribution well-defined.

◆ getNumberOfThreads()

Size gum::learning::CachedContingencyCounter::getNumberOfThreads ( ) const

overridevirtualinherited

returns the current max number of threads of the scheduler

Implements gum::IThreadNumberManager.

◆ isGumNumberOfThreadsOverriden()

bool gum::learning::CachedContingencyCounter::isGumNumberOfThreadsOverriden ( ) const

overridevirtualinherited

indicates whether the user set herself the number of threads

Implements gum::IThreadNumberManager.

◆ marginalize_()

std::vector< double > gum::learning::IndependenceTest::marginalize_	(	const std::size_t	node_2_marginalize,
		const std::size_t	X_size,
		const std::size_t	Y_size,
		const std::size_t	Z_size,
		const std::vector< double > &	N_xyz ) const

protected

returns a counting vector where variables are marginalized from N_xyz

Parameters

node_2_marginalize	indicates which node(s) shall be marginalized: 0 means that X should be marginalized 1 means that Y should be marginalized 2 means that Z should be marginalized
X_size	the domain size of variable X
Y_size	the domain size of variable Y
Z_size	the domain size of the set of conditioning variables Z
N_xyz	a counting vector of dimension X * Y * Z (in this order)

◆ minNbRowsPerThread()

virtual std::size_t gum::learning::CachedContingencyCounter::minNbRowsPerThread ( ) const

virtualinherited

returns the minimum of rows that each thread should process

◆ nodeId2Columns()

const Bijection< NodeId, std::size_t > & gum::learning::CachedContingencyCounter::nodeId2Columns ( ) const

inherited

return the mapping between the columns of the database and the node ids

Warning: An empty nodeId2Columns bijection means that the mapping is an identity, i.e., the value of a NodeId is equal to the index of the column in the DatabaseTable.

◆ operator=() [1/2]

IndependenceTest & gum::learning::IndependenceTest::operator= ( const IndependenceTest & from )

protected

copy operator

References IndependenceTest().

Here is the call graph for this function:

◆ operator=() [2/2]

IndependenceTest & gum::learning::IndependenceTest::operator= ( IndependenceTest && from )

protected

move operator

References IndependenceTest().

Here is the call graph for this function:

◆ ranges()

const std::vector< std::pair< std::size_t, std::size_t > > & gum::learning::CachedContingencyCounter::ranges ( ) const

inherited

returns the current ranges

Referenced by CachedContingencyCounter(), gum::learning::IndependenceTest::IndependenceTest(), gum::learning::IndepTestChi2::IndepTestChi2(), gum::learning::IndepTestG2::IndepTestG2(), and gum::learning::KNML::KNML().

Here is the caller graph for this function:

◆ setMinNbRowsPerThread()

virtual void gum::learning::CachedContingencyCounter::setMinNbRowsPerThread ( const std::size_t nb ) const

virtualinherited

changes the number min of rows a thread should process in a multithreading context

◆ setNumberOfThreads()

void gum::learning::CachedContingencyCounter::setNumberOfThreads ( Size nb )

overridevirtualinherited

sets the number max of threads that can be used

Implements gum::IThreadNumberManager.

◆ setRanges()

void gum::learning::CachedContingencyCounter::setRanges ( const std::vector< std::pair< std::size_t, std::size_t > > & new_ranges )

inherited

sets new ranges to perform the counts

◆ statistics()

virtual std::pair< double, double > gum::learning::IndependenceTest::statistics	(	NodeId	var1,
		NodeId	var2,
		const std::vector< NodeId > &	rhs_ids = {} )

pure virtual

returns the pair (test statistic, p-value) for the independence test X _|_ Y | Z

Implemented in gum::learning::IndepTestChi2, and gum::learning::IndepTestG2.

◆ useCache()

virtual void gum::learning::CachedContingencyCounter::useCache ( const bool on_off )

virtualinherited

turn on/off the use of a cache of the previously computed score

Reimplemented in gum::learning::KNML.

Member Data Documentation

◆ _domain_sizes_

std::vector< std::size_t > gum::learning::IndependenceTest::_domain_sizes_

protected

the domain sizes of the variables (indexed by column id in the database)

Definition at line 148 of file independenceTest.h.

◆ cache_

ScoringCache gum::learning::CachedContingencyCounter::cache_

protectedinherited

the scoring cache

Definition at line 171 of file cachedContingencyCounter.h.

◆ counter_

RecordCounter gum::learning::CachedContingencyCounter::counter_

protectedinherited

the record counter used for the counts over discrete variables

Definition at line 168 of file cachedContingencyCounter.h.

◆ empty_ids_

const std::vector< NodeId > gum::learning::CachedContingencyCounter::empty_ids_

protectedinherited

an empty vector

Definition at line 177 of file cachedContingencyCounter.h.

◆ one_log2_

const double gum::learning::CachedContingencyCounter::one_log2_ {M_LOG2E}

protectedinherited

1 / log(2)

Definition at line 162 of file cachedContingencyCounter.h.

162{M_LOG2E};

M_LOG2E

#define M_LOG2E

Definition math_utils.h:55

◆ prior_

Prior* gum::learning::CachedContingencyCounter::prior_ {nullptr}

protectedinherited

the expert knowledge prior added to the contingency tables

Definition at line 165 of file cachedContingencyCounter.h.

165{nullptr};

◆ use_cache_

bool gum::learning::CachedContingencyCounter::use_cache_ {true}

protectedinherited

a Boolean indicating whether we wish to use the cache

Definition at line 174 of file cachedContingencyCounter.h.

174{true};

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

agrum/base/stattests/independenceTest.h

Public Member Functions

Protected Member Functions

Static Protected Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

◆ IndependenceTest() [1/4]

◆ IndependenceTest() [2/4]

◆ ~IndependenceTest()

◆ IndependenceTest() [3/4]

◆ IndependenceTest() [4/4]

Member Function Documentation

◆ clear()

◆ clearCache()

◆ clearRanges()

◆ clone()

◆ computeStatistics_()

◆ database()

◆ degreesOfFreedom_()

◆ getNumberOfThreads()

◆ isGumNumberOfThreadsOverriden()

◆ marginalize_()

◆ minNbRowsPerThread()

◆ nodeId2Columns()

◆ operator=() [1/2]

◆ operator=() [2/2]

◆ ranges()

◆ setMinNbRowsPerThread()

◆ setNumberOfThreads()

◆ setRanges()

◆ statistics()

◆ useCache()

Member Data Documentation

◆ _domain_sizes_

◆ cache_

◆ counter_

◆ empty_ids_

◆ one_log2_

◆ prior_

◆ use_cache_