keel.Algorithms.Neural_Networks.ensemble

Class Ensemble

java.lang.Object

keel.Algorithms.Neural_Networks.ensemble.Ensemble

public class Ensemble
extends java.lang.Object

Class representing an ensemble

Since:

JDK1.5

Version:

0.1

Author:

Written by Nicolas Garcia Pedrajas (University of Cordoba) 27/02/2007

Field Summary

Fields

Modifier and Type	Field and Description
double[]	betta Ensemble weights.
double[][][][]	cache Ensemble weights.
int	TEST Training flag.
int	TRAIN Training flag.
int	VAL Training flag.
double[][]	weights Ensemble weights.

Constructor Summary

Constructors

Constructor and Description
Ensemble(EnsembleParameters global) Constructor

Method Summary

Modifier and Type	Method and Description
void	EnsembleOutput(double[] inputs, double[] outputs) Output of every network
void	GetAdaWeights() Calculate weights using Ada method
void	GetGEMWeights(EnsembleParameters global, double[][] data, int n) Calculate weights using GEM method
void	LoadEnsemble(java.lang.String file_name) Load ensemble from file_name
void	SaveEnsemble(java.lang.String file_name, java.lang.String header) Save ensemble at file_name
void	SaveOutputFile(java.lang.String file_name, double[][] data, int n, java.lang.String problem, double[] a, double[] b) Save data in output file
double	TestEnsembleInClassification(EnsembleParameters global, double[][] data, int npatterns) Test ensemble in classification
double	TestEnsembleInRegression(EnsembleParameters global, double[][] data, int npatterns) Test ensemble in regression
void	TrainEnsemble(EnsembleParameters global, Data data) Train Ensemble
void	TrainEnsembleAda(EnsembleParameters global, Data data) Train ensemble using Ada
void	TrainEnsembleArcing(EnsembleParameters global, Data data) Train ensemble using Arcing
void	TrainEnsembleBagging(EnsembleParameters global, Data data) Train ensemble using Bagging
void	TrainEnsembleNoSampling(EnsembleParameters global, Data data) Train ensemble without sampling
void	UpdateCache(EnsembleParameters global, Data data) Update cache data

Methods inherited from class java.lang.Object

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

Field Detail

weights

public double[][] weights

Ensemble weights.

betta

public double[] betta

Ensemble weights.

cache

public double[][][][] cache

Ensemble weights.

TRAIN

public final int TRAIN

Training flag.

See Also:

Constant Field Values

TEST

public final int TEST

Training flag.

See Also:

Constant Field Values

VAL

public final int VAL

Training flag.

See Also:

Constant Field Values

Constructor Detail

Ensemble

public Ensemble(EnsembleParameters global)

Constructor

Parameters:

global - Global definition parameters

Method Detail

TrainEnsemble

public void TrainEnsemble(EnsembleParameters global,
                          Data data)

Train Ensemble

Parameters:

global - Global definition parameters

data - Input data

TrainEnsembleNoSampling

public void TrainEnsembleNoSampling(EnsembleParameters global,
                                    Data data)

Train ensemble without sampling

Parameters:

global - Global definition parameters

data - Input data

TrainEnsembleBagging

public void TrainEnsembleBagging(EnsembleParameters global,
                                 Data data)

Train ensemble using Bagging

Parameters:

global - Global definition parameters

data - Input data

TrainEnsembleArcing

public void TrainEnsembleArcing(EnsembleParameters global,
                                Data data)

Train ensemble using Arcing

Parameters:

global - Global definition parameters

data - Input data.

TrainEnsembleAda

public void TrainEnsembleAda(EnsembleParameters global,
                             Data data)

Train ensemble using Ada

Parameters:

global - Global definition parameters

data - Input data

EnsembleOutput

public void EnsembleOutput(double[] inputs,
                           double[] outputs)

Output of every network

Parameters:

inputs - Input data

outputs - Output data

TestEnsembleInClassification

public double TestEnsembleInClassification(EnsembleParameters global,
                                           double[][] data,
                                           int npatterns)

Test ensemble in classification

Parameters:

global - Global definition parameters

data - Input data

npatterns - No of patterns

Returns:

Correct classified per cent

TestEnsembleInRegression

public double TestEnsembleInRegression(EnsembleParameters global,
                                       double[][] data,
                                       int npatterns)

Test ensemble in regression

Parameters:

global - Global definition parameters

data - Input data

npatterns - No of patterns

Returns:

Test ensemble in regression

SaveEnsemble

public void SaveEnsemble(java.lang.String file_name,
                         java.lang.String header)

Save ensemble at file_name

Parameters:

file_name - File name

header - header of the data set for which the network has been adjusted to

LoadEnsemble

public void LoadEnsemble(java.lang.String file_name)

Load ensemble from file_name

Parameters:

file_name - File name

GetGEMWeights

public void GetGEMWeights(EnsembleParameters global,
                          double[][] data,
                          int n)

Calculate weights using GEM method

Parameters:

global - Global definition parameters

data - Input data

n - matrix order (no of rows and colms in data matrix)

GetAdaWeights

public void GetAdaWeights()

Calculate weights using Ada method

SaveOutputFile

public void SaveOutputFile(java.lang.String file_name,
                           double[][] data,
                           int n,
                           java.lang.String problem,
                           double[] a,
                           double[] b)

Save data in output file

Parameters:

file_name - File name

data - Data to be saved

n - No of patterns

problem - Type of problem (CLASSIFICATION | REGRESSION)

a - Scaling parameter (a).

b - Scaling parameter (b).

UpdateCache

public void UpdateCache(EnsembleParameters global,
                        Data data)

Update cache data

Parameters:

global - Global definition parameters

data - Input data