Package edu.cmu.tetrad.search

Class Ling

java.lang.Object

edu.cmu.tetrad.search.Ling

public class Ling
extends Object

The code used within this class is largely Gustave Lacerda's, which corresponds to his essay, Discovering Cyclic Causal Models by Independent Components Analysis. The code models the LiNG algorithm.

Note: This code is currently broken; please do not use it until it's fixed. 11/24/2015> 0

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	Ling.StoredGraphs	This small class is used to store graph permutations.

Constructor Summary

Constructors

Constructor	Description
Ling(DataSet d)	The algorithm only requires a DataSet to process.
Ling(Graph g, int samples)	When you don't have a Dataset, supply a Graph and the number of samples to draw and the algorithm will generate a DataSet.
Ling(GraphWithParameters graphWP, int samples)	When you don't have a Dataset, supply a GraphWithParameters and the number of samples to draw and the algorithm will generate a DataSet.

Method Summary

Modifier and Type	Method	Description
DataSet	getData()
long	getElapsedTime()	Calculates the time used when processing the search method.
double	ngFullData(int rowIndex, Matrix data, Matrix W)
Matrix	pruneEdgesByResampling(Matrix data)	This is the method used in Patrik's code.
Ling.StoredGraphs	search()	The search method is used to process LiNG.
void	setThreshold(double t)	Sets the value at which thresholding occurs on Fast ICA data.

Methods inherited from class java.lang.Object

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

Constructor Details

Ling

public Ling(DataSet d)

The algorithm only requires a DataSet to process. Passing in a Dataset and then running the search algorithm is an effetive way to use LiNG.

Parameters:

d - a DataSet over which the algorithm can process

Ling

public Ling(GraphWithParameters graphWP,
 int samples)

When you don't have a Dataset, supply a GraphWithParameters and the number of samples to draw and the algorithm will generate a DataSet.

Parameters:

graphWP - a graph with parameters from GraphWithParameters

samples - the number of samples the algorithm draws in order to generate a DataSet

Ling

public Ling(Graph g,
 int samples)

When you don't have a Dataset, supply a Graph and the number of samples to draw and the algorithm will generate a DataSet.

Parameters:

g - a graph from Graph

samples - the number of samples the algorithm draws in order to generate a DataSet

Method Details

getData

public DataSet getData()

Returns:

DataSet Returns a dataset of the data used by the algorithm.

search

public Ling.StoredGraphs search()

The search method is used to process LiNG. Call search when you want to run the algorithm.

ngFullData

public double ngFullData(int rowIndex,
 Matrix data,
 Matrix W)

getElapsedTime

public long getElapsedTime()

Calculates the time used when processing the search method.

setThreshold

public void setThreshold(double t)

Sets the value at which thresholding occurs on Fast ICA data. Default is .05.

Parameters:

t - The value at which the thresholding is set

pruneEdgesByResampling

public Matrix pruneEdgesByResampling(Matrix data)

This is the method used in Patrik's code.