Package edu.cmu.tetrad.search.kernel

Class KernelUtils

java.lang.Object

edu.cmu.tetrad.search.kernel.KernelUtils

public class KernelUtils
extends Object

Various kernel utilities

Author:

Robert Tillman

Constructor Summary

Constructors

Constructor	Description
KernelUtils()

Method Summary

Modifier and Type	Method	Description
static Matrix	constructCentralizedGramMatrix(List<Kernel> kernels, DataSet dataset, List<Node> nodes)	Constructs the centralized Gram matrix for a given vector valued sample.
static Matrix	constructGramMatrix(List<Kernel> kernels, DataSet dataset, List<Node> nodes)	Constructs Gram matrix for a given vector valued sample.
static Matrix	constructH(int m)	Constructs the projection matrix on 1/m
static Matrix	incompleteCholeskyGramMatrix(List<Kernel> kernels, DataSet dataset, List<Node> nodes, double precision)	Approximates Gram matrix using incomplete Cholesky factorization

Methods inherited from class java.lang.Object

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

Constructor Details

KernelUtils

public KernelUtils()

Method Details

constructGramMatrix

public static Matrix constructGramMatrix(List<Kernel> kernels,
 DataSet dataset,
 List<Node> nodes)

Constructs Gram matrix for a given vector valued sample. The set of kernels corresponds to the variables in the set. The output matrix is the tensor product of Gram matrices for each variable.

Parameters:

kernels - the kernels for each variable

dataset - the dataset containing each variable

nodes - the variables to construct the Gram matrix for

constructCentralizedGramMatrix

public static Matrix constructCentralizedGramMatrix(List<Kernel> kernels,
 DataSet dataset,
 List<Node> nodes)

Constructs the centralized Gram matrix for a given vector valued sample. The set of kernels corresponds to the variables in the set. The output matrix is the tensor product of Gram matrices for each variable.

Parameters:

kernels - the kernels for each variable

dataset - the dataset containing each variable

nodes - the variables to construct the Gram matrix for

constructH

public static Matrix constructH(int m)

Constructs the projection matrix on 1/m

Parameters:

m - the sample size

incompleteCholeskyGramMatrix

public static Matrix incompleteCholeskyGramMatrix(List<Kernel> kernels,
 DataSet dataset,
 List<Node> nodes,
 double precision)

Approximates Gram matrix using incomplete Cholesky factorization

Parameters:

kernels - the kernels for each variable

dataset - the dataset containing each variable

nodes - the variables to construct the Gram matrix for