Module es.upm.fi.cig.multictbnc

Package es.upm.fi.cig.multictbnc.performance

Class HoldOutMethod

java.lang.Object

es.upm.fi.cig.multictbnc.performance.ValidationMethod

es.upm.fi.cig.multictbnc.performance.HoldOutMethod

public class HoldOutMethod
extends ValidationMethod

Implements hold-out validation method.

Constructor Summary

Constructors

Constructor	Description
HoldOutMethod(DatasetReader datasetReader, double trainingSize, boolean estimateProbabilities, boolean shuffle, Long seed)	Constructs a HoldOut by receiving a DatasetReader, the size of the training set and if the data should be shuffled.

Method Summary

Modifier and Type	Method	Description
Map<String,Double>	evaluate(MultiCTBNC<?,?> model)	Evaluates the performance of the specified model using hold-out validation.
Map<String,Double>	evaluate(MultiCTBNC<?,?> model, double preprocessingExecutionTime)	Evaluates the performance of the specified model and returns the results.
void	generateTrainAndTest()	Generates a training and a test dataset.
Dataset	getTraining()	Returns the training dataset.

Methods inherited from class es.upm.fi.cig.multictbnc.performance.ValidationMethod

displayModel, displayResults, setWriter

Methods inherited from class java.lang.Object

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

Constructor Details

HoldOutMethod

public HoldOutMethod(DatasetReader datasetReader,
 double trainingSize,
 boolean estimateProbabilities,
 boolean shuffle,
 Long seed)

Constructs a HoldOut by receiving a DatasetReader, the size of the training set and if the data should be shuffled.

Parameters:

datasetReader - a DatasetReader to read the dataset

trainingSize - size of the training dataset (percentage)

estimateProbabilities - true to estimate the probabilities of the class configurations, false otherwise

shuffle - true to shuffle the sequences before splitting them into training and test datasets, false otherwise

seed - seed used to shuffle the sequences

Method Details

generateTrainAndTest

public void generateTrainAndTest()
                          throws UnreadDatasetException

Generates a training and a test dataset.

Throws:

UnreadDatasetException - if the provided dataset could not be read

getTraining

public Dataset getTraining()

Returns the training dataset.

Returns:

training dataset

evaluate

public Map<String,Double> evaluate(MultiCTBNC<?,?> model)
                            throws UnreadDatasetException,
ErroneousValueException

Evaluates the performance of the specified model using hold-out validation.

Specified by:

evaluate in class ValidationMethod

Parameters:

model - model to evaluate

Returns:

test results in a Map

Throws:

UnreadDatasetException - if the provided dataset could not be read

ErroneousValueException - if a provided parameter is erroneous for the requested task

evaluate

public Map<String,Double> evaluate(MultiCTBNC<?,?> model,
 double preprocessingExecutionTime)
                            throws UnreadDatasetException,
ErroneousValueException

Description copied from class: ValidationMethod

Evaluates the performance of the specified model and returns the results. This method receives the execution time of previous tasks so it is add up to the total mean learning time.

Specified by:

evaluate in class ValidationMethod

Parameters:

model - model to evaluate

preprocessingExecutionTime - execution time of previous tasks

Returns:

results saved in a Map.

Throws:

UnreadDatasetException - if there was an error while reading a dataset

ErroneousValueException - if a provided parameter is erroneous for the requested task