Package edu.cmu.tetrad.search

Class Mimbuild

java.lang.Object

edu.cmu.tetrad.search.Mimbuild

public class Mimbuild
extends Object

An implemetation of Mimbuild based on the Fgsl score. The search will attempt a GES search first and if that throws and exception then a CPC search. The penalty penaltyDiscount parameter is for the GES search; the alpha value is for the CPC search. Or you can just grab the latent covariance matrix and run whatever search you want to. (I don't know why GES sometimes fails, it is a mystery.)

Uses a different (better) algorithm from Mimbuild. Preferable.

Author:

Joseph Ramsey

Constructor Summary

Constructors

Constructor	Description
Mimbuild()

Method Summary

Modifier and Type	Method	Description
List<List<Node>>	getClustering()
double	getEpsilon()
Graph	getFullGraph()
Knowledge	getKnowledge()
List<String>	getLatentNames(List<Node> latents)
ICovarianceMatrix	getLatentsCov()
int	getMinClusterSize()	jf Clusters smaller than this size will be tossed out.
double	getMinimum()
int	getNumParams()
double	getpValue()
Graph	search(List<List<Node>> clustering, List<String> latentNames, ICovarianceMatrix measuresCov)
void	setEpsilon(double epsilon)	Parameter convergence threshold.
void	setKnowledge(Knowledge knowledge)
void	setMinClusterSize(int minClusterSize)
void	setPenaltyDiscount(double penaltyDiscount)

Methods inherited from class java.lang.Object

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

Constructor Details

Mimbuild

public Mimbuild()

Method Details

search

public Graph search(List<List<Node>> clustering,
 List<String> latentNames,
 ICovarianceMatrix measuresCov)

getClustering

public List<List<Node>> getClustering()

getKnowledge

public Knowledge getKnowledge()

setKnowledge

public void setKnowledge(Knowledge knowledge)

getLatentsCov

public ICovarianceMatrix getLatentsCov()

getLatentNames

public List<String> getLatentNames(List<Node> latents)

getMinimum

public double getMinimum()

getpValue

public double getpValue()

getFullGraph

public Graph getFullGraph()

Returns:

the allowUnfaithfulness discovered graph, with latents and indicators.

getEpsilon

public double getEpsilon()

setEpsilon

public void setEpsilon(double epsilon)

Parameter convergence threshold. Default = 1e-4.

setPenaltyDiscount

public void setPenaltyDiscount(double penaltyDiscount)

getNumParams

public int getNumParams()

getMinClusterSize

public int getMinClusterSize()

jf Clusters smaller than this size will be tossed out.

setMinClusterSize

public void setMinClusterSize(int minClusterSize)