org.neuroph.contrib.bpbench

Class BackPropBenchmarks

java.lang.Object

org.neuroph.contrib.bpbench.BackPropBenchmarks

public class BackPropBenchmarks
extends java.lang.Object

Base class for benchmarking of backpropagation algorithms. Defines list of trainings and number of repetitions of each treaning.

Constructor Summary

Constructors

Constructor and Description
BackPropBenchmarks() Create an instance with empty list of training`s
BackPropBenchmarks(java.util.List<AbstractTraining> listOfTasks) Create an instance with given list of training`s

Method Summary

Modifier and Type	Method and Description
void	addTraining(AbstractTraining training) Adds new training to list
int	getNoOfRepetitions() Returns number of repetitions for each training
void	run() Executes all training`s from list with predefined number of repetitions and resets neural net
void	saveResults(java.lang.String filePath) Save result of benchmarking in given location as csv
void	setNoOfRepetitions(int noOfRepetitions) Adds number of repetitions
void	startBenchmark(java.util.List<java.lang.Class<? extends AbstractTraining>> trainingTypeCollection, java.util.List<TrainingSettings> settingsCollection, DataSet trainingSet) Creates all training`s using list training types, settings and execute run method
void	startBenchmark(java.util.List<java.lang.Class<? extends AbstractTraining>> trainingTypeCollection, java.util.List<TrainingSettings> settingsCollection, DataSet trainingSet, MultiLayerPerceptron mlp) Creates all training`s using list training types, settings and neural network and execute run method

Methods inherited from class java.lang.Object

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

Constructor Detail

BackPropBenchmarks

public BackPropBenchmarks()

Create an instance with empty list of training`s

BackPropBenchmarks

public BackPropBenchmarks(java.util.List<AbstractTraining> listOfTasks)

Create an instance with given list of training`s

Parameters:

listOfTasks - list of training`s

Method Detail

addTraining

public void addTraining(AbstractTraining training)

Adds new training to list

Parameters:

training -

run

public void run()

Executes all training`s from list with predefined number of repetitions and resets neural net

getNoOfRepetitions

public int getNoOfRepetitions()

Returns number of repetitions for each training

Returns:

number of repetitions

setNoOfRepetitions

public void setNoOfRepetitions(int noOfRepetitions)

Adds number of repetitions

Parameters:

noOfRepetitions -

startBenchmark

public void startBenchmark(java.util.List<java.lang.Class<? extends AbstractTraining>> trainingTypeCollection,
                           java.util.List<TrainingSettings> settingsCollection,
                           DataSet trainingSet,
                           MultiLayerPerceptron mlp)

Creates all training`s using list training types, settings and neural network and execute run method

Parameters:

trainingTypeCollection -

settingsCollection -

trainingSet -

mlp -

startBenchmark

public void startBenchmark(java.util.List<java.lang.Class<? extends AbstractTraining>> trainingTypeCollection,
                           java.util.List<TrainingSettings> settingsCollection,
                           DataSet trainingSet)

Creates all training`s using list training types, settings and execute run method

Parameters:

trainingTypeCollection -

settingsCollection -

trainingSet -

saveResults

public void saveResults(java.lang.String filePath)

Save result of benchmarking in given location as csv

Parameters:

filePath -