Package fr.lip6.jkernelmachines.evaluation

Provides interface and implementations for evaluation metrics (e.g. accuracy, average precision) and cross-validation.

See: Description

Interface Summary

Interface	Description
BalancedCrossValidation	Interface for crossvalidation classes that can balance the number of sample per class while computing the splits.
CrossValidation	Standard interface for all cross-validation methods
Evaluator<T>	Basic interface for all evaluation tools.
MultipleEvaluatorCrossValidation<T>

Class Summary

Class	Description
AccuracyEvaluator<T>	Simple evaluation class for computing the accuracy on a testing set.
ApEvaluator<T>	Simple evaluation class for computing the mean average precision, VOC style.
Evaluation<T>	Simple class containing a sample and its evaluation by the classifier
FScoreEvaluator<T>	Evaluator computing the F-score (by default F1).
LeaveOneOutCrossValidation<T>	Simple class to perform leave one out cross validation
MulticlassAccuracyEvaluator<T>	Evaluation class for computing the multiclass accuracy on a testing set, given a provided cmulticlass classifier.
MultipleEvaluator<T>	Class that aggregates several Evaluators
NFoldCrossValidation<T>	Class for performing N-Fold Cross-validation.
PrecisionEvaluator<T>	Evaluator computing the precision defined as (number of relevant)/(number of retrieved).
RandomSplitCrossValidation<T>	Class for simple random split based cross validation.

Package fr.lip6.jkernelmachines.evaluation Description

Provides interface and implementations for evaluation metrics (e.g. accuracy, average precision) and cross-validation.