Package edu.cmu.tetrad.search.score

Class ConditionalGaussianScore

java.lang.Object

edu.cmu.tetrad.search.score.ConditionalGaussianScore

All Implemented Interfaces:

Score

public class ConditionalGaussianScore
extends Object
implements Score

Implements a conditional Gaussian BIC score for FGS, which calculates a BIC score for mixed discrete/Gaussian data using the conditional Gaussian likelihood function (see). The reference is here:

Andrews, B., Ramsey, J., & Cooper, G. F. (2018). Scoring Bayesian networks of mixed variables. International journal of data science and analytics, 6, 3-18.

As for all scores in Tetrad, higher scores mean more dependence, and negative scores indicate independence.

Version:

$Id: $Id

Author:

josephramsey

See Also:

• ConditionalGaussianLikelihood
• DegenerateGaussianScore

Constructor Summary

Constructors

Constructor	Description
ConditionalGaussianScore(DataSet dataSet, double penaltyDiscount, boolean discretize)	Constructs the score using a covariance matrix.

Method Summary

Modifier and Type	Method	Description
int	getMaxDegree()	Returns the max degree, by default 1000.
double	getPenaltyDiscount()	Returns the penalty discount for this score, which is a multiplier on the penalty term of the BIC score.
int	getSampleSize()	Returns the sample size of the data.
List<Node>	getVariables()	The variables of the score.
boolean	isEffectEdge(double bump)	Returns true iff the edge between x and y is an effect edge.
double	localScore(int i, int... parents)	Calculates the sample likelihood and BIC score for index i given its parents in a simple SEM model.
double	localScoreDiff(int x, int y, int[] z)	Calculates localScore(y | z, x) - localScore(z).
void	setNumCategoriesToDiscretize(int numCategoriesToDiscretize)	Sets tne number of categories used to discretize, when this optimization is used.
void	setPenaltyDiscount(double penaltyDiscount)	Sets the penalty discount for this score, which is a multiplier on the penalty discount of the BIC score.
void	setStructurePrior(double structurePrior)	Setter for the field structurePrior.
String	toString()	A string representation of the score.

Methods inherited from class java.lang.Object

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

Methods inherited from interface edu.cmu.tetrad.search.score.Score

append, determines, getVariable, localScore, localScore, localScoreDiff

Constructor Details

ConditionalGaussianScore

public ConditionalGaussianScore(DataSet dataSet,
 double penaltyDiscount,
 boolean discretize)

Constructs the score using a covariance matrix.

Parameters:

dataSet - A dataset with a mixture of continuous and discrete variables. It may be all continuous or all discrete.

penaltyDiscount - A multiplier on the penalty term in the BIC score.

discretize - When a discrete variable is a child of a continuous variable, one (expensive) way to solve the problem is to do a numerical integration. A less expensive (and often more accurate) way to solve the problem is to discretize the child with a certain number of discrete categories. if this parameter is set to True, a separate copy of all variables is maintained that is discretized in this way, and these are substituted for the discrete children when this sort of problem needs to be solved. This information needs to be known in the constructor since one needs to know right away whether ot create this separate discretized version of the continuous columns.

See Also:

• setNumCategoriesToDiscretize(int)

Method Details

localScore

public double localScore(int i,
 int... parents)

Calculates the sample likelihood and BIC score for index i given its parents in a simple SEM model.

Specified by:

localScore in interface Score

Parameters:

i - The index of the child.

parents - The indices of the parents.

Returns:

The score.,

localScoreDiff

public double localScoreDiff(int x,
 int y,
 int[] z)

Calculates localScore(y | z, x) - localScore(z).

Specified by:

localScoreDiff in interface Score

Parameters:

x - The index of the child.

y - a int

z - The indices of the parents.

Returns:

The score difference.

getSampleSize

public int getSampleSize()

Returns the sample size of the data.

Specified by:

getSampleSize in interface Score

Returns:

This size.

isEffectEdge

public boolean isEffectEdge(double bump)

Returns true iff the edge between x and y is an effect edge.

A method for FGES for determining whether an edge counts as an effect edges for this score bump.

Specified by:

isEffectEdge in interface Score

Parameters:

bump - The bump.

Returns:

True iff the edge between x and y is an effect edge.

See Also:

• Fges

getVariables

public List<Node> getVariables()

The variables of the score.

Returns the variables of the data.

Specified by:

getVariables in interface Score

Returns:

This list.

getMaxDegree

public int getMaxDegree()

Returns the max degree, by default 1000.

Returns the max degree recommended for the search form the MagSemBicScore and Fges.

Specified by:

getMaxDegree in interface Score

Returns:

The max degree.

See Also:

• MagSemBicScore
• Fges

getPenaltyDiscount

public double getPenaltyDiscount()

Returns the penalty discount for this score, which is a multiplier on the penalty term of the BIC score.

Returns:

This penalty discount.

setPenaltyDiscount

public void setPenaltyDiscount(double penaltyDiscount)

Sets the penalty discount for this score, which is a multiplier on the penalty discount of the BIC score.

Parameters:

penaltyDiscount - This penalty discount.

setNumCategoriesToDiscretize

public void setNumCategoriesToDiscretize(int numCategoriesToDiscretize)

Sets tne number of categories used to discretize, when this optimization is used.

Parameters:

numCategoriesToDiscretize - This number.

toString

public String toString()

A string representation of the score.

Specified by:

toString in interface Score

Overrides:

toString in class Object

Returns:

This string.

setStructurePrior

public void setStructurePrior(double structurePrior)

Setter for the field structurePrior.

Parameters:

structurePrior - a double