de.lmu.ifi.dbs.elki.algorithm.itemsetmining.associationrules.interest

Class GiniIndex

java.lang.Object

de.lmu.ifi.dbs.elki.algorithm.itemsetmining.associationrules.interest.GiniIndex

All Implemented Interfaces:

InterestingnessMeasure

@Reference(authors="P. Tan, V. Kumar",title="Interestingness measures for association patterns: A perspective",booktitle="Proc. Workshop on Postprocessing in Machine Learning and Data Mining",url="https://www.cs.umn.edu/sites/cs.umn.edu/files/tech_reports/00-036.pdf",bibkey="tr/umn/TanK00") @Reference(authors="L. Breiman, J. H. Friedman, R. A. Olshen, C. J. Stone",title="Classification and Regression Trees",booktitle="",bibkey="books/wa/BreimanFOS84")
public class GiniIndex
extends java.lang.Object
implements InterestingnessMeasure

Gini-index based interestingness measure, using the weighted squared conditional probabilities compared to the non-conditional priors.

\[ P(X)\left(P(Y|X)^2+P(\neg Y|X)^2\right) + P(\neg X)\left(P(Y|\neg X)^2+P(\neg Y|\neg X)^2\right) - P(Y)^2 - P(\neg Y)^2 \]

Reference:

P. Tan, V. Kumar
Interestingness measures for association patterns: A perspective
In Proc. Workshop on Postprocessing in Machine Learning and Data Mining

The Gini was originally used in decision trees:

L. Breiman, J. H. Friedman, R. A. Olshen, C. J. Stone
Classification and Regression Trees

Since:

0.7.5

Author:

Frederic Sautter

Constructor Summary

Constructors

Constructor and Description
GiniIndex() Constructor.

Method Summary

Modifier and Type	Method and Description
double	measure(int t, int sX, int sY, int sXY) Computes the value of the measure for a given support values

Methods inherited from class java.lang.Object

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

Constructor Detail

GiniIndex

public GiniIndex()

Constructor.

Method Detail

measure

public double measure(int t,
                      int sX,
                      int sY,
                      int sXY)

Description copied from interface: InterestingnessMeasure

Computes the value of the measure for a given support values

Specified by:

measure in interface InterestingnessMeasure

Parameters:

t - Total number of transaction

sX - Support of the antecedent

sY - Support of the consequent

sXY - Support of the union of antecedent and consequent

Returns:

value of the measure