Package weka.core.metrics

Interface Summary
InstanceConverter InstanceConverter interface Certain metrics may require converting instances before performing computations
 

Class Summary
AttrEvalMetricLearner AttrEvalMetricLearner - sets the weights of a metric using scores from an attribute evaluator
BarHillelMetric Class for performing RCA according to Bar-Hillel's algorithm.
BarHillelMetricMatlab Class for performing RCA according to Bar-Hillel's algorithm.
ClassifierMetricLearner ClassifierMetricLearner - learns metric parameters by constructing "difference instances" and then learning weights that classify same-class instances as positive, and different-class instances as negative.
GDMetricLearner GDMetricLearner - sets the weights of a metric using gradient descent
HardPairwiseSelector HardPairwiseSelector class.
KL KL class Implements weighted Kullback-Leibler divergence
LearnableMetric Interface to distance metrics that can be learned
MatlabMetricLearner MatlabMetricLearner - learns metric parameters by constructing "difference instances" and then learning weights that classify same-class instances as positive, and different-class instances as negative using an external Matlab program.
Metric Abstract Metric class
MetricLearner Abstract MetricLearner interface.
PairwiseSelector Abstract PairwiseSelector class.
RandomPairwiseSelector RandomPairwiseSelector class.
TrainingPair This is a basic class for a distance training pair
WeightedDotP WeightedDotP class Implements the weighted dot product distance metric
WeightedEuclidean WeightedEuclidean class Implements weighted euclidean distance metric
WeightedMahalanobis WeightedMahalanobis class Implements a weighted Mahalanobis distance metric weighted by a full matrix of weights.
XingMetric Class for performing RCA according to Bar-Hillel's algorithm.