Package edu.cmu.tetrad.search

Class GFci

java.lang.Object

edu.cmu.tetrad.search.GFci

All Implemented Interfaces:

IGraphSearch

public final class GFci
extends Object
implements IGraphSearch

Implements a modification of FCI that started by running the FGES algorithm and then fixes that result to be correct for latent variables models. First, colliders from the FGES results are copied into the final circle-circle graph, and some independence reasoning is used to add the remaining colliders into the graph. Then, the FCI final orientation rules are applied. The reference is here:

Ogarrio, J. M., Spirtes, P., & Ramsey, J. (2016, August). A hybrid causal search algorithm for latent variable models. In Conference on probabilistic graphical models (pp. 368-379). PMLR.

Because the method both runs FGES (a score-based algorithm) and does additional checking of conditional independencies, both as part of its collider orientation step and also as part of the the definite discriminating path step in the final FCI orientation rules, both a score and a test need to be used to construct a GFCI algorithm.

Note that various score-based algorithms could be used in place of FGES for the initial step; in this repository we give three other options, GRaSP-FCI, BFCI (BOSS FCI), and SP-FCI (see).

For more information on the algorithm, see the reference above.

This class is configured to respect knowledge of forbidden and required edges, including knowledge of temporal tiers.

Version:

$Id: $Id

Author:

Juan Miguel Ogarrio, peterspirtes, josephramsey

See Also:

• Fci
• FciOrient
• GraspFci
• BFci
• SpFci
• Fges
• Knowledge

Constructor Summary

Constructors

Constructor	Description
GFci(IndependenceTest test, Score score)	Constructs a new GFci algorithm with the given independence test and score.

Method Summary

Modifier and Type	Method	Description
IndependenceTest	getIndependenceTest()	Returns the independence test used in search.
Knowledge	getKnowledge()	Returns the knowledge used in search.
Graph	search()	Runs the graph and returns the search PAG.
void	setCompleteRuleSetUsed(boolean completeRuleSetUsed)	Sets whether Zhang's complete rules are used.
void	setDepth(int depth)	Sets the depth of the search for the possible m-sep search.
void	setFaithfulnessAssumed(boolean faithfulnessAssumed)	Sets whether one-edge faithfulness is assumed.
void	setGuaranteePag(boolean guaranteePag)	Sets the flag indicating whether to guarantee the output is a legal PAG.
void	setKnowledge(Knowledge knowledge)	Sets the knowledge to use in search.
void	setMaxDegree(int maxDegree)	Sets the maximum indegree of the output graph.
void	setMaxDiscriminatingPathLength(int maxDiscriminatingPathLength)	Sets the maximum length of any discriminating path.
void	setNumThreads(int numThreads)	Sets the number of threads to use in the search.
void	setOut(PrintStream out)	Sets the print stream used for output, default System.out.
void	setSepsetFinderMethod(int sepsetFinderMethod)	Sets the method used to find the sepset in the GFci algorithm.
void	setVerbose(boolean verbose)	Sets whether verbose output should be printed.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

GFci

public GFci(IndependenceTest test,
 Score score)

Constructs a new GFci algorithm with the given independence test and score.

Parameters:

test - The independence test to use.

score - The score to use.

Method Details

search

public Graph search()
             throws InterruptedException

Runs the graph and returns the search PAG.

Specified by:

search in interface IGraphSearch

Returns:

This PAG.

Throws:

InterruptedException - if any.

setMaxDegree

public void setMaxDegree(int maxDegree)

Sets the maximum indegree of the output graph.

Parameters:

maxDegree - This maximum.

getKnowledge

public Knowledge getKnowledge()

Returns the knowledge used in search.

Returns:

This knowledge

setKnowledge

public void setKnowledge(Knowledge knowledge)

Sets the knowledge to use in search.

Parameters:

knowledge - This knowledge.

setCompleteRuleSetUsed

public void setCompleteRuleSetUsed(boolean completeRuleSetUsed)

Sets whether Zhang's complete rules are used.

Parameters:

completeRuleSetUsed - set to true if Zhang's complete rule set should be used, false if only R1-R4 (the rule set of the original FCI) should be used. True by default.

setMaxDiscriminatingPathLength

public void setMaxDiscriminatingPathLength(int maxDiscriminatingPathLength)

Sets the maximum length of any discriminating path.

Parameters:

maxDiscriminatingPathLength - the maximum length of any discriminating path, or -1 if unlimited.

setVerbose

public void setVerbose(boolean verbose)

Sets whether verbose output should be printed.

Parameters:

verbose - True, if so.

getIndependenceTest

public IndependenceTest getIndependenceTest()

Returns the independence test used in search.

Returns:

This test.

setOut

public void setOut(PrintStream out)

Sets the print stream used for output, default System.out.

Parameters:

out - This print stream.

setFaithfulnessAssumed

public void setFaithfulnessAssumed(boolean faithfulnessAssumed)

Sets whether one-edge faithfulness is assumed. For FGES

Parameters:

faithfulnessAssumed - True, if so.

See Also:

• Fges.setFaithfulnessAssumed(boolean)

setDepth

public void setDepth(int depth)

Sets the depth of the search for the possible m-sep search.

Parameters:

depth - This depth.

setNumThreads

public void setNumThreads(int numThreads)

Sets the number of threads to use in the search.

Parameters:

numThreads - The number of threads to use. Must be at least 1.

setGuaranteePag

public void setGuaranteePag(boolean guaranteePag)

Sets the flag indicating whether to guarantee the output is a legal PAG.

Parameters:

guaranteePag - A boolean value indicating whether to guarantee the output is a legal PAG.

setSepsetFinderMethod

public void setSepsetFinderMethod(int sepsetFinderMethod)

Sets the method used to find the sepset in the GFci algorithm.

Parameters:

sepsetFinderMethod - The method used to find the sepset. - 0: Default method - 1: Custom method 1 - 2: Custom method 2 - ...