org.android.activityminer.logic.evaluation
Class CrossValidationStrategy

java.lang.Object
  org.android.activityminer.logic.evaluation.CrossValidationStrategy

All Implemented Interfaces:

ValidationStrategy

public final class CrossValidationStrategy
extends Object
implements ValidationStrategy

Implementation off the k-fold cross validation based evaluation strategy.

Field Summary

private double	avgPredictionTime The average prediction time.
private double	avgTrainingTime The average training time per sample.
private long[]	classIds The set to store the available activity classes in
private Set<Set<TrainingActivity>>	evaluationSets The equal sized evaluation sets to be used for training and test set creation.
private int	k The cross validation fold count
private double	maxPredictionTime The maximum prediction time for a sample.
private double	minPredictionTime The minimum prediction time for a sample.

Constructor Summary

CrossValidationStrategy(int k)
Constructor

Method Summary

EvaluationResult	eval(Set<FeatureTypes> visibleAttributes, Classifier classifier) Does evaluate the classification results for the given classifier and visible attributes.
double	getAvgPreditionTime() Getter for the average time to predict an activity
double	getAvgTrainingTime() Getter for the average time to train an activity sample.
String	getDescription() Getter for the evaluation strategy description.
double	getMaxPreditionTime() Getter for the maximum prediction time
double	getMinPreditionTime() Getter for the minimum prediction time
void	prepare(List<TrainingActivity> trainingActivities) Method for preparation with the set of training activities.
boolean	validateEvaluationSet() Validation method for evaluation set initialization.

Methods inherited from class java.lang.Object

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

Field Detail

k

private final int k

The cross validation fold count

avgTrainingTime

private double avgTrainingTime

The average training time per sample.

avgPredictionTime

private double avgPredictionTime

The average prediction time.

minPredictionTime

private double minPredictionTime

The minimum prediction time for a sample.

maxPredictionTime

private double maxPredictionTime

The maximum prediction time for a sample.

evaluationSets

private Set<Set<TrainingActivity>> evaluationSets

The equal sized evaluation sets to be used for training and test set creation.

classIds

private long[] classIds

The set to store the available activity classes in

Constructor Detail

CrossValidationStrategy

public CrossValidationStrategy(int k)

Constructor

Parameters:

k - the cross validation fold count

Method Detail

getDescription

public final String getDescription()

Description copied from interface: ValidationStrategy

Getter for the evaluation strategy description.

Specified by:

getDescription in interface ValidationStrategy

Returns:

the description of the evaluation strategy.

prepare

public final void prepare(List<TrainingActivity> trainingActivities)

Description copied from interface: ValidationStrategy

Method for preparation with the set of training activities.

Specified by:

prepare in interface ValidationStrategy

Parameters:

trainingActivities - the set of training activities.

eval

public final EvaluationResult eval(Set<FeatureTypes> visibleAttributes,
                                   Classifier classifier)

Description copied from interface: ValidationStrategy

Does evaluate the classification results for the given classifier and visible attributes.

Specified by:

eval in interface ValidationStrategy

Parameters:

visibleAttributes - the visible attributes for evaluation

classifier - the classifier to use for evaluation

Returns:

the evaluation result

validateEvaluationSet

public final boolean validateEvaluationSet()

Validation method for evaluation set initialization.

Returns:

true if set is initialized with data

getAvgPreditionTime

public double getAvgPreditionTime()

Description copied from interface: ValidationStrategy

Getter for the average time to predict an activity

Specified by:

getAvgPreditionTime in interface ValidationStrategy

Returns:

the average time to predict an activity

getAvgTrainingTime

public double getAvgTrainingTime()

Description copied from interface: ValidationStrategy

Getter for the average time to train an activity sample.

Specified by:

getAvgTrainingTime in interface ValidationStrategy

Returns:

the average time to train an activity sample

getMaxPreditionTime

public double getMaxPreditionTime()

Description copied from interface: ValidationStrategy

Getter for the maximum prediction time

Specified by:

getMaxPreditionTime in interface ValidationStrategy

Returns:

the maximum time to predict an activity

getMinPreditionTime

public double getMinPreditionTime()

Description copied from interface: ValidationStrategy

Getter for the minimum prediction time

Specified by:

getMinPreditionTime in interface ValidationStrategy

Returns:

the minimum time to predict an activity