de.datexis.sector.eval

Class ClassificationScoreCalculator

java.lang.Object

org.deeplearning4j.earlystopping.scorecalc.base.BaseIEvaluationScoreCalculator<org.deeplearning4j.nn.api.Model,ClassificationEvaluation>

de.datexis.sector.eval.ClassificationScoreCalculator

All Implemented Interfaces:

Serializable, org.deeplearning4j.earlystopping.scorecalc.ScoreCalculator<org.deeplearning4j.nn.api.Model>

public class ClassificationScoreCalculator
extends org.deeplearning4j.earlystopping.scorecalc.base.BaseIEvaluationScoreCalculator<org.deeplearning4j.nn.api.Model,ClassificationEvaluation>

Score function for evaluating a MultiLayerNetwork according to an evaluation metric (Evaluation.Metric such as accuracy, F1 score, etc. Used for both MultiLayerNetwork and ComputationGraph

Author:

Alex Black

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected LookupCacheEncoder	encoder
protected Tagger	tagger

Fields inherited from class org.deeplearning4j.earlystopping.scorecalc.base.BaseIEvaluationScoreCalculator

iter, iterator

Constructor Summary

Constructors

Constructor and Description
ClassificationScoreCalculator(Tagger tagger, LookupCacheEncoder encoder, org.nd4j.linalg.dataset.api.iterator.DataSetIterator iterator)
ClassificationScoreCalculator(Tagger tagger, LookupCacheEncoder encoder, org.nd4j.linalg.dataset.api.iterator.MultiDataSetIterator iterator)

Method Summary

Modifier and Type	Method and Description
double	calculateScore(org.deeplearning4j.nn.api.Model network)
protected void	evaluate(org.deeplearning4j.nn.graph.ComputationGraph net, ClassificationEvaluation evaluation, org.nd4j.linalg.dataset.api.iterator.MultiDataSetIterator iterator) Override evaluation to use average of forward/backward layers in a single score.
protected double	finalScore(ClassificationEvaluation e)
boolean	minimizeScore()
protected ClassificationEvaluation	newEval()

Methods inherited from class java.lang.Object

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

Field Detail

tagger

protected Tagger tagger

encoder

protected LookupCacheEncoder encoder

Constructor Detail

ClassificationScoreCalculator

public ClassificationScoreCalculator(Tagger tagger,
                                     LookupCacheEncoder encoder,
                                     org.nd4j.linalg.dataset.api.iterator.DataSetIterator iterator)

ClassificationScoreCalculator

public ClassificationScoreCalculator(Tagger tagger,
                                     LookupCacheEncoder encoder,
                                     org.nd4j.linalg.dataset.api.iterator.MultiDataSetIterator iterator)

Method Detail

newEval

protected ClassificationEvaluation newEval()

Specified by:

newEval in class org.deeplearning4j.earlystopping.scorecalc.base.BaseIEvaluationScoreCalculator<org.deeplearning4j.nn.api.Model,ClassificationEvaluation>

calculateScore

public double calculateScore(org.deeplearning4j.nn.api.Model network)

Specified by:

calculateScore in interface org.deeplearning4j.earlystopping.scorecalc.ScoreCalculator<org.deeplearning4j.nn.api.Model>

Overrides:

calculateScore in class org.deeplearning4j.earlystopping.scorecalc.base.BaseIEvaluationScoreCalculator<org.deeplearning4j.nn.api.Model,ClassificationEvaluation>

evaluate

protected void evaluate(org.deeplearning4j.nn.graph.ComputationGraph net,
                        ClassificationEvaluation evaluation,
                        org.nd4j.linalg.dataset.api.iterator.MultiDataSetIterator iterator)

Override evaluation to use average of forward/backward layers in a single score.

finalScore

protected double finalScore(ClassificationEvaluation e)

Specified by:

finalScore in class org.deeplearning4j.earlystopping.scorecalc.base.BaseIEvaluationScoreCalculator<org.deeplearning4j.nn.api.Model,ClassificationEvaluation>

minimizeScore

public boolean minimizeScore()