org.knime.base.node.preproc.discretization.caim2.modelcreator
Class QuantaMatrix2D

java.lang.Object
  org.knime.base.node.preproc.discretization.caim2.modelcreator.QuantaMatrix2D

public class QuantaMatrix2D
extends Object

Implements a 2D quanta matrix that holds for one dimension class values and for the other dimension the intervals of discretization.

Author:

Christoph Sieb, University of Konstanz

Constructor Summary

QuantaMatrix2D(DiscretizationScheme dScheme, Map<String,Integer> classValueToIndexMap)
Initializes a new 2D quanta matrix for class value - interval statistics.

Method Summary

double	calculateCaim() Calculates the current caim value for the quanta matrix.
void	countData(BufferedDataTable inputTable, int columnIndex, int classColumnIndex) Counts the distribution of the given column of the input table and stores it into the quanta matrix.

Methods inherited from class java.lang.Object

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

Constructor Detail

QuantaMatrix2D

public QuantaMatrix2D(DiscretizationScheme dScheme,
                      Map<String,Integer> classValueToIndexMap)

Initializes a new 2D quanta matrix for class value - interval statistics.

Parameters:

dScheme - the discretization scheme representing the first dimension

classValueToIndexMap - the class values representing the other dimension contains a mapping from the name to the index used for the matrix values

Method Detail

countData

public void countData(BufferedDataTable inputTable,
                      int columnIndex,
                      int classColumnIndex)

Counts the distribution of the given column of the input table and stores it into the quanta matrix.

Parameters:

inputTable - the table to create the statistics for

columnIndex - the index of the column this matrix should be used for

classColumnIndex - the index of the class column of the input table

calculateCaim

public double calculateCaim()

Calculates the current caim value for the quanta matrix.

Returns:

the caim value as defined by Kurgan et al.