com.itemanalysis.psychometrics.polycor

Class CovarianceMatrix

java.lang.Object

com.itemanalysis.psychometrics.polycor.CovarianceMatrix

public class CovarianceMatrix
extends java.lang.Object

Computes NxN covariance and correlation matrix. The Covariance class will automatically do a pairwise deletion. For listwise deletion, the data should be screened for missing values before calling increment() and increment should only be called when a particular case has data for all variables. For example, do not call increment() if a case is missing data on any of the variables in a matrix and listwise deletion is required.

Since:

January 27, 2008

Author:

J. Patrick Meyer

Constructor Summary

Constructors

Constructor and Description
CovarianceMatrix(java.util.ArrayList<VariableAttributes> variables)
CovarianceMatrix(java.util.ArrayList<VariableAttributes> variables, boolean unbiased)
CovarianceMatrix(Covariance[][] covMat, java.util.ArrayList<VariableAttributes> variables)
CovarianceMatrix(Covariance[][] covMat, java.util.ArrayList<VariableAttributes> variables, boolean unbiased)
CovarianceMatrix(double[][] covarianceMatrix)
CovarianceMatrix(double[][] covarianceMatrix, boolean unbiased) Input is a two-array of doubles.
CovarianceMatrix(int numberOfVariables)
CovarianceMatrix(int numberOfVariables, boolean unbiased) This constructor is primarily used in the TestSummary.java class.
CovarianceMatrix(java.util.LinkedHashMap<VariableName,VariableAttributes> variableAttributeMap)
CovarianceMatrix(java.util.LinkedHashMap<VariableName,VariableAttributes> variableAttributeMap, boolean unbiased)

Method Summary

Modifier and Type	Method and Description
double	columnSum(int columnIndex)
double[][]	correlation()
double	covarianceSum()
double	diagonalSum()
double	getCorrelationAt(int i, int j)
double	getCovarianceAt(int i, int j)
double	getMaxSampleSize()
int	getNumberOfVariables()
double	getPvalueAt(int i, int j)
double	getStandardErrorAt(int i, int j)
double	getVarianceAt(int index)
void	increment(int xIndex, int yIndex, java.lang.Double x, java.lang.Double y) Increment values in the matrix.
double	listwiseSampleSize()
java.lang.String	printCorrelationMatrix(boolean showStdError)
java.lang.String	printCovarianceMatrix()
double	rowSum(int rowIndex)
void	setNameAt(int index, java.lang.String name)
void	setNameAt(int index, VariableName name)
java.lang.String	toString()
double	totalVariance()
double[][]	value()

Methods inherited from class java.lang.Object

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

Constructor Detail

CovarianceMatrix

public CovarianceMatrix(double[][] covarianceMatrix,
                        boolean unbiased)

Input is a two-array of doubles. Each entry is a covariance that may or may not have been computed as an unbiased estimate. The unbiased argument here is somewhat misleading and probably unnecessary.

Parameters:

covarianceMatrix - a two-way array of doubles that are covariances

unbiased -

CovarianceMatrix

public CovarianceMatrix(double[][] covarianceMatrix)

CovarianceMatrix

public CovarianceMatrix(java.util.ArrayList<VariableAttributes> variables,
                        boolean unbiased)

CovarianceMatrix

public CovarianceMatrix(java.util.ArrayList<VariableAttributes> variables)

CovarianceMatrix

public CovarianceMatrix(java.util.LinkedHashMap<VariableName,VariableAttributes> variableAttributeMap,
                        boolean unbiased)

CovarianceMatrix

public CovarianceMatrix(java.util.LinkedHashMap<VariableName,VariableAttributes> variableAttributeMap)

CovarianceMatrix

public CovarianceMatrix(int numberOfVariables,
                        boolean unbiased)

This constructor is primarily used in the TestSummary.java class.

Parameters:

numberOfVariables -

CovarianceMatrix

public CovarianceMatrix(int numberOfVariables)

CovarianceMatrix

public CovarianceMatrix(Covariance[][] covMat,
                        java.util.ArrayList<VariableAttributes> variables,
                        boolean unbiased)

CovarianceMatrix

public CovarianceMatrix(Covariance[][] covMat,
                        java.util.ArrayList<VariableAttributes> variables)

Method Detail

setNameAt

public void setNameAt(int index,
                      VariableName name)

setNameAt

public void setNameAt(int index,
                      java.lang.String name)

getMaxSampleSize

public double getMaxSampleSize()

increment

public void increment(int xIndex,
                      int yIndex,
                      java.lang.Double x,
                      java.lang.Double y)

Increment values in the matrix. Only increments for the upper diagonal of the matrix.

Parameters:

xIndex -

yIndex -

x -

y -

value

public double[][] value()

correlation

public double[][] correlation()

getStandardErrorAt

public double getStandardErrorAt(int i,
                                 int j)

getPvalueAt

public double getPvalueAt(int i,
                          int j)

getCovarianceAt

public double getCovarianceAt(int i,
                              int j)

getCorrelationAt

public double getCorrelationAt(int i,
                               int j)

totalVariance

public double totalVariance()

diagonalSum

public double diagonalSum()

covarianceSum

public double covarianceSum()

getVarianceAt

public double getVarianceAt(int index)

rowSum

public double rowSum(int rowIndex)

columnSum

public double columnSum(int columnIndex)

getNumberOfVariables

public int getNumberOfVariables()

listwiseSampleSize

public double listwiseSampleSize()

Returns:

double indicating sample size. This method is only valid for listwise deletion. Sample sizes vary for each pair with pairwise deletion.

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object

printCorrelationMatrix

public java.lang.String printCorrelationMatrix(boolean showStdError)

printCovarianceMatrix

public java.lang.String printCovarianceMatrix()