Package edu.cmu.tetrad.search

Class Lofs

java.lang.Object

edu.cmu.tetrad.search.Lofs

public class Lofs
extends Object

Implements a number of methods which take a fixed graph as input and use linear, non-Gaussian methods to orient the edges in the graph. where the acronym stands for linear, non-Gaussian Orientation with a Fixed graph Structure (LOFS). The options for different types of scores are given in the enum Lofs.Score. The options for rules to use to do the orientations are given in the enum, Lofs.Rule. Most of these are taken from the literature and can be googled, though we should certainly give this reference for several of them, to which we are indebted:

Hyvärinen, A., & Smith, S. M. (2013). Pairwise likelihood ratios for estimation of non-Gaussian structural equation models. The Journal of Machine Learning Research, 14(1), 111-152.

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

Version:

$Id: $Id

Author:

josephramsey

See Also:

• Lofs.Score
• Lofs.Rule
• Knowledge

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static enum	Lofs.Rule	Give a list of options for rules for doing the non-Gaussian orientations.
static enum	Lofs.Score	Gives a list of options for non-Gaussian transformations that can be used for some scores.

Constructor Summary

Constructors

Constructor	Description
Lofs(Graph graph, List<DataSet> dataSets)	Constructor.

Method Summary

Modifier and Type	Method	Description
double	getPValue(double fisherZ)	Calculates the p-value using Fisher's exact test.
double	kernel(double z)	Calculates the kernel value for a given input.
double	kernel1(double z)	Computes the value of the kernel function 1 for a given input.
double	kernel2(double z)	Calculates the value of the kernel2 function for the given input.
double	kernel3(double z)	This method calculates the value of the kernel function, kernel3, for a given input.
double	kernel4(double z)	Calculates the result of kernel4 function.
double	kernel5(double z)	Calculates the value of the kernel function, kernel5, for the given input.
double	kernel6(double z)	Calculates the value of the kernel function 6 given a parameter z.
double	kernel7(double z)	Computes the value of the kernel7 function for the given value.
double	kernel8(double z)	Calculates the value of the kernel8 function for the given input.
Graph	orient()	Orients the graph and returns the oriented graph.
double	scoreRow(int rowIndex, Matrix data, List<List<Integer>> rows, List<List<Double>> parameters)	Calculates the score for a given row in a matrix using the specified parameters.
void	setAlpha(double alpha)	Sets the alpha to use, where applicable.
void	setEpsilon(double epsilon)	Sets the epsilon value.
void	setKnowledge(Knowledge knowledge)	Sets the knowledge for the object.
void	setOrientStrongerDirection(boolean orientStrongerDirection)	Sets whether orientation should be done in the stronger direction, where applicable.
void	setR2Orient2Cycles(boolean r2Orient2Cycles)	Sets for R2 whether cycles should be oriented.
void	setRule(Lofs.Rule rule)	Sets the rule to use to do the orientation.
void	setScore(Lofs.Score score)	Sets the (LoFS) score to use.
void	setSelfLoopStrength(double selfLoopStrength)	Sets the self-loop strength, if applicable.

Methods inherited from class java.lang.Object

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

Constructor Details

Lofs

public Lofs(Graph graph,
 List<DataSet> dataSets)
     throws IllegalArgumentException

Constructor.

Parameters:

graph - The graph to be oriented. Orientations for the graph will be overwritten.

dataSets - A list of datasets to use to do the orientation. This may be just one dataset. If more than one dataset are given, the data will be concatenated (pooled).

Throws:

IllegalArgumentException - if any.

Method Details

orient

public Graph orient()

Orients the graph and returns the oriented graph.

Returns:

The oriented graph.

setRule

public void setRule(Lofs.Rule rule)

Sets the rule to use to do the orientation.

Parameters:

rule - This rule.

See Also:

• Lofs.Rule

setScore

public void setScore(Lofs.Score score)

Sets the (LoFS) score to use.

Parameters:

score - This score.

See Also:

• Lofs.Score

setSelfLoopStrength

public void setSelfLoopStrength(double selfLoopStrength)

Sets the self-loop strength, if applicable.

Parameters:

selfLoopStrength - This strength.

setAlpha

public void setAlpha(double alpha)

Sets the alpha to use, where applicable.

Parameters:

alpha - This alpha.

setOrientStrongerDirection

public void setOrientStrongerDirection(boolean orientStrongerDirection)

Sets whether orientation should be done in the stronger direction, where applicable.

Parameters:

orientStrongerDirection - True, if so.

setR2Orient2Cycles

public void setR2Orient2Cycles(boolean r2Orient2Cycles)

Sets for R2 whether cycles should be oriented.

Parameters:

r2Orient2Cycles - True, if so.

scoreRow

public double scoreRow(int rowIndex,
 Matrix data,
 List<List<Integer>> rows,
 List<List<Double>> parameters)

Calculates the score for a given row in a matrix using the specified parameters.

Parameters:

rowIndex - the index of the row to score

data - the matrix containing the data

rows - the list of rows containing the column indices to score

parameters - the list of parameters for each column index in the rows

Returns:

the score of the row

setEpsilon

public void setEpsilon(double epsilon)

Sets the epsilon value.

Parameters:

epsilon - the new value of epsilon

setKnowledge

public void setKnowledge(Knowledge knowledge)

Sets the knowledge for the object.

Parameters:

knowledge - the knowledge to set.

Throws:

NullPointerException - if the knowledge is null.

getPValue

public double getPValue(double fisherZ)

Calculates the p-value using Fisher's exact test. The p-value is the probability of observing a test statistic as extreme as the observed value under the null hypothesis. It is used to determine the statistical significance of the observed value.

Parameters:

fisherZ - the Fisher's Z value

Returns:

the p-value

kernel

public double kernel(double z)

Calculates the kernel value for a given input.

Parameters:

z - the input value to calculate the kernel value for.

Returns:

the calculated kernel value for the given input.

kernel1

public double kernel1(double z)

Computes the value of the kernel function 1 for a given input. (Gaussian.)

Parameters:

z - the input value for which the kernel function is computed

Returns:

the value of the kernel function computed for the given input value

kernel2

public double kernel2(double z)

Calculates the value of the kernel2 function for the given input. (Uniform.)

Parameters:

z - the input value to calculate the kernel2 for

Returns:

the calculated value of kernel2

kernel3

public double kernel3(double z)

This method calculates the value of the kernel function, kernel3, for a given input. (Triangular.)

Parameters:

z - the input value for which the kernel function is to be calculated

Returns:

the calculated value of the kernel function

kernel4

public double kernel4(double z)

Calculates the result of kernel4 function. (Epanechnikov)

Parameters:

z - a double value representing the input parameter

Returns:

the result of kernel4 function

kernel5

public double kernel5(double z)

Calculates the value of the kernel function, kernel5, for the given input. (Quartic)

The kernel function calculates the value based on the input value z. If the absolute value of z is greater than 1, the function returns 0. Otherwise, it returns the result of the calculation: (15 / 16)

Parameters:

z - the input parameter

Returns:

the result of kernel5 function

kernel6

public double kernel6(double z)

Calculates the value of the kernel function 6 given a parameter z. (Triweight)

Parameters:

z - the input parameter

Returns:

the calculated value of the kernel function 6

kernel7

public double kernel7(double z)

Computes the value of the kernel7 function for the given value. (Tricube)

Parameters:

z - the input value

Returns:

the computed value of the kernel7 function

kernel8

public double kernel8(double z)

Calculates the value of the kernel8 function for the given input. (Cosine)

Parameters:

z - the input value for the kernel8 function

Returns:

the calculated value of the kernel8 function