keel.Algorithms.Discretizers.Bayesian_Discretizer

Class BayesianDiscretizer

java.lang.Object

keel.Algorithms.Discretizers.Basic.Discretizer

keel.Algorithms.Discretizers.Bayesian_Discretizer.BayesianDiscretizer

public class BayesianDiscretizer
extends Discretizer

This is the class with the operations of the Bayesian discretization. It adopts the behavior of the general discretizers and specifies its differences in this class, that has to extend the abstract methods.

Since:

JDK1.5

Version:

1.0

Author:

Written by Victoria Lopez Morales (University of Granada) 21/12/2009

Field Summary

Fields inherited from class keel.Algorithms.Discretizers.Basic.Discretizer

classOfInstances, cutPoints, iClassIndex, realAttributes, realValues

Constructor Summary

Constructors

Constructor and Description
BayesianDiscretizer()

Method Summary

Modifier and Type	Method and Description
protected java.util.Vector	discretizeAttribute(int attribute, int[] values, int begin, int end) Selects, for a given attribute, the real values that best discretize the attribute according to the bayesian discretizer

Methods inherited from class keel.Algorithms.Discretizers.Basic.Discretizer

applyDiscretization, buildCutPoints, discretize, getCutPoint, getNumIntervals, sortValues

Methods inherited from class java.lang.Object

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

Constructor Detail

BayesianDiscretizer

public BayesianDiscretizer()

Method Detail

discretizeAttribute

protected java.util.Vector discretizeAttribute(int attribute,
                                               int[] values,
                                               int begin,
                                               int end)

Selects, for a given attribute, the real values that best discretize the attribute according to the bayesian discretizer

Specified by:

discretizeAttribute in class Discretizer

Parameters:

attribute - Position of the attribute in the list of attributes whose best real values for discretization are going to be selected

values - Position of the corresponding attribute value in the real values matrix, ordered by attribute value

begin - First value that is considered to belong to the data considered, usually 0

end - Last value that is considered to belong to the data considered, usually the last value of the dataset

Returns:

a vector with the real values that best discretize the attribute given according to the bayesian discretizer