meka.experiment.evaluators

Interface Evaluator

All Superinterfaces:

LogSupporter, weka.core.OptionHandler, Serializable

All Known Implementing Classes:

AbstractEvaluator, AbstractMetaEvaluator, CrossValidation, PercentageSplit, RepeatedRuns

public interface Evaluator
extends weka.core.OptionHandler, Serializable, LogSupporter

Interface for classes that evaluate on a dataset.

Version:

$Revision$

Author:

FracPete (fracpete at waikato dot ac dot nz)

Method Summary

Modifier and Type	Method and Description
List<EvaluationStatistics>	evaluate(MultiLabelClassifier classifier, weka.core.Instances dataset) Returns the evaluation statistics generated for the dataset.
String	initialize() Initializes the evaluator.
void	stop() Stops the evaluation, if possible.

Methods inherited from interface weka.core.OptionHandler

getOptions, listOptions, setOptions

Methods inherited from interface meka.events.LogSupporter

addLogListener, log, removeLogListener

Method Detail

initialize

String initialize()

Initializes the evaluator.

Returns:

null if successfully initialized, otherwise error message

evaluate

List<EvaluationStatistics> evaluate(MultiLabelClassifier classifier,
                                    weka.core.Instances dataset)

Returns the evaluation statistics generated for the dataset.

Parameters:

classifier - the classifier to evaluate

dataset - the dataset to evaluate on

Returns:

the statistics

stop

void stop()

Stops the evaluation, if possible.