keel.Algorithms.Lazy_Learning.IDIBL

Class IDIBL

java.lang.Object

keel.Algorithms.Lazy_Learning.LazyAlgorithm

keel.Algorithms.Lazy_Learning.IDIBL.IDIBL

public class IDIBL
extends LazyAlgorithm

File: IDIBL.java The Integrated Decremental Instance Based Learning algorithm. A complete IBL classifier, wich uses IVDM distance, Drop4 prunning algorithm, and a tunning algorithm to optimize the parameters used in the final K-nn classification process

Since:

JDK1.5

Version:

1.0

Author:

Written by Joaquín Derrac (University of Granada) 16/11/2008

Field Summary

Fields inherited from class keel.Algorithms.Lazy_Learning.LazyAlgorithm

confMatrix, initialTime, inputAtt, inputs, modelTime, name, nClasses, nInstances, nulls, outFile, output, prediction, predictions, realClass, reference, referenceData, referenceFile, referenceOutput, relation, seed, temp, test, testData, testFile, testOutput, testTime, train, trainConfMatrix, trainData, trainFile, trainingTime, trainOutput, trainPrediction, trainRealClass, trainUnclassified, unclassified

Constructor Summary

Constructors

Constructor and Description
IDIBL(java.lang.String script) The main method of the class

Method Summary

Modifier and Type	Method and Description
protected int	evaluate(double[] example) Evaluates a instance to predict its class.
void	findNeigbours() Finds MAXK Nearest Neighbors for every reference instance
void	findParameters(int timeLimit) Finds best parameters for drop 4 algorithm
void	findProbabilities() Find probabilities to calculate IVDM distances
int	getFirstTimeLimit() Get first tunning algorithm time limit
int	getSecondTimeLimit() Get second tunning algorithm time limit
void	pruneInstanceSet() Prune instance set using drop 4 algorithm
protected void	readParameters(java.lang.String script) Reads configuration script, to extract the parameter's values.
void	reTuneParameters(int timeLimit) Finds best parameters for classification

Methods inherited from class keel.Algorithms.Lazy_Learning.LazyAlgorithm

euclideanDistance, evaluate2, execute, executeReference, manhattanDistance, normalizeReference, normalizeTest, normalizeTrain, printInstance, readConfiguracion, readDataFiles, readDataFiles, readDataFiles, same, setInitialTime

Methods inherited from class java.lang.Object

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

Constructor Detail

IDIBL

public IDIBL(java.lang.String script)

The main method of the class

Parameters:

script - Name of the configuration script

Method Detail

readParameters

protected void readParameters(java.lang.String script)

Reads configuration script, to extract the parameter's values.

Specified by:

readParameters in class LazyAlgorithm

Parameters:

script - Name of the configuration script

evaluate

protected int evaluate(double[] example)

Evaluates a instance to predict its class.

Specified by:

evaluate in class LazyAlgorithm

Parameters:

example - Instance evaluated

Returns:

Class selected

findProbabilities

public void findProbabilities()

Find probabilities to calculate IVDM distances

findNeigbours

public void findNeigbours()

Finds MAXK Nearest Neighbors for every reference instance

findParameters

public void findParameters(int timeLimit)

Finds best parameters for drop 4 algorithm

Parameters:

timeLimit - Limit of iterations

pruneInstanceSet

public void pruneInstanceSet()

Prune instance set using drop 4 algorithm

reTuneParameters

public void reTuneParameters(int timeLimit)

Finds best parameters for classification

Parameters:

timeLimit - Limit of iterations

getFirstTimeLimit

public int getFirstTimeLimit()

Get first tunning algorithm time limit

Returns:

Time limit

getSecondTimeLimit

public int getSecondTimeLimit()

Get second tunning algorithm time limit

Returns:

Time limit