edu.illinois.cs.cogcomp.core.experiments

Class ClassificationTester

java.lang.Object

edu.illinois.cs.cogcomp.core.experiments.ClassificationTester

public class ClassificationTester
extends Object

A class for recording results of experiments. This can be used to generate a result table with one row per label and a confusion table.

Author:

Vivek Srikumar

Constructor Summary

Constructors

Constructor and Description
ClassificationTester()

Method Summary

Modifier and Type	Method and Description
Pair<Table,List<String>>	getConfusionTable()
EvaluationRecord	getEvaluationRecord()
EvaluationRecord	getEvaluationRecord(String label)
Set<String>	getLabels()
double	getMacroF1()
double	getMacroPrecision()
double	getMacroRecall()
double	getMicroF1()
double	getMicroPrecision()
double	getMicroRecall()
Table	getPerformanceTable()
Table	getPerformanceTable(boolean printCounts)
void	ignoreLabelFromSummary(String label)
void	record(String goldLabel, String predictedLabel)
void	recordCount(String label, int goldCount, int predictedCount, int correctCount)
void	recordGoldOnly(String goldLabel)
void	recordPredictionOnly(String predictedLabel)
void	reset()
void	setSignificanceInfo(Map<String,ShufflingBasedStatisticalSignificance> significance) Statistical significance information.

Methods inherited from class java.lang.Object

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

Constructor Detail

ClassificationTester

public ClassificationTester()

Method Detail

ignoreLabelFromSummary

public void ignoreLabelFromSummary(String label)

reset

public void reset()

getEvaluationRecord

public EvaluationRecord getEvaluationRecord()

getEvaluationRecord

public EvaluationRecord getEvaluationRecord(String label)

getLabels

public Set<String> getLabels()

recordCount

public void recordCount(String label,
                        int goldCount,
                        int predictedCount,
                        int correctCount)

record

public void record(String goldLabel,
                   String predictedLabel)

recordGoldOnly

public void recordGoldOnly(String goldLabel)

recordPredictionOnly

public void recordPredictionOnly(String predictedLabel)

getConfusionTable

public Pair<Table,List<String>> getConfusionTable()

getMicroF1

public double getMicroF1()

getMicroPrecision

public double getMicroPrecision()

getMicroRecall

public double getMicroRecall()

getMacroF1

public double getMacroF1()

getMacroPrecision

public double getMacroPrecision()

getMacroRecall

public double getMacroRecall()

getPerformanceTable

public Table getPerformanceTable()

getPerformanceTable

public Table getPerformanceTable(boolean printCounts)

setSignificanceInfo

public void setSignificanceInfo(Map<String,ShufflingBasedStatisticalSignificance> significance)

Statistical significance information. Takes map from label to p-values for {P, R, F} for that label.