Examples of DifferenceToReferenceLanguageBCR

statechum.analysis.learning.experiments.PairSelection.PairQualityLearner.DifferenceToReferenceLanguageBCR

Examples of statechum.analysis.learning.experiments.PairSelection.PairQualityLearner.DifferenceToReferenceLanguageBCR

      LearnerEvaluationConfiguration learnerEval = new LearnerEvaluationConfiguration(config);learnerEval.setLabelConverter(converter);
      int states = referenceGraph.getAcceptStateNumber();
      final Collection<List<Label>> testSet = PaperUAS.computeEvaluationSet(referenceGraph,states,PairQualityLearner.makeEven(states*referenceGraph.pathroutines.computeAlphabet().size()));
      
      DifferenceToReferenceDiff differenceStructural=DifferenceToReferenceDiff.estimationOfDifferenceDiffMeasure(referenceGraph, learntGraph, config, 1);
      DifferenceToReferenceLanguageBCR differenceBCRlearnt=DifferenceToReferenceLanguageBCR.estimationOfDifference(referenceGraph, learntGraph,testSet);
    
      final MarkovModel m= new MarkovModel(chunkLen,true,true,false);
      LearnerGraph pta=new LearnerGraph(config);
      for(List<Label> seq:sPlus)
        pta.paths.augmentPTA(seq,true,false,null);
      for(List<Label> seq:sMinus)
        pta.paths.augmentPTA(seq,false,false,null);
      pta.clearColours();
      new MarkovClassifier(m, pta).updateMarkov(false);// construct Markov chain
      // For Markov, we do not need to learn anything at all - our Markov matrix contains enough information to classify paths and hence compare it to the reference graph.
      ConfusionMatrix mat = DiffExperiments.classifyAgainstMarkov(testSet, referenceGraph, m);
      DifferenceToReferenceLanguageBCR differenceBCRMarkov = new DifferenceToReferenceLanguageBCR(mat);
      
      return new OtpErlangTuple(new OtpErlangObject[]{
          new OtpErlangDouble(differenceStructural.getValue()),new OtpErlangDouble(differenceBCRlearnt.getValue()),new OtpErlangDouble(differenceBCRMarkov.getValue())
      });
    }

View Full Code Here

Examples of statechum.analysis.learning.experiments.PairSelection.PairQualityLearner.DifferenceToReferenceLanguageBCR

      }


      {// For Markov, we do not need to learn anything at all - our Markov matrix contains enough information to classify paths and hence compare it to the reference graph.
        ConfusionMatrix mat = DiffExperiments.classifyAgainstMarkov(testSet, referenceGraph, m);
        dataSample.markovLearner = new ScoresForGraph();          
        dataSample.markovLearner.differenceBCR = new DifferenceToReferenceLanguageBCR(mat);
      }
      
      dataSample.fractionOfStatesIdentifiedBySingletons=Math.round(100*MarkovClassifier.calculateFractionOfStatesIdentifiedBySingletons(referenceGraph));
      dataSample.stateNumber = referenceGraph.getStateNumber();
      dataSample.transitionsSampled = Math.round(100*trimmedReference.pathroutines.countEdges()/referenceGraph.pathroutines.countEdges());

View Full Code Here

Examples of statechum.analysis.learning.experiments.PairSelection.PairQualityLearner.DifferenceToReferenceLanguageBCR


        {// For Markov, we do not need to learn anything at all - our Markov matrix contains enough information to classify paths and hence compare it to the reference graph.
          ConfusionMatrix mat = DiffExperiments.classifyAgainstMarkov(testSet, referenceGraph, m);
          dataSample.markovLearner = new ScoresForGraph();      
          System.out.println("Markov");
          dataSample.markovLearner.differenceBCR = new DifferenceToReferenceLanguageBCR(mat);
          System.out.println(dataSample.markovLearner.differenceBCR.getValue());
        }
        
        dataSample.fractionOfStatesIdentifiedBySingletons=Math.round(100*MarkovClassifier.calculateFractionOfStatesIdentifiedBySingletons(referenceGraph));
        dataSample.stateNumber = referenceGraph.getStateNumber();

View Full Code Here

Examples of statechum.analysis.learning.experiments.PairSelection.PairQualityLearner.DifferenceToReferenceLanguageBCR

      LearnerEvaluationConfiguration learnerEval = new LearnerEvaluationConfiguration(config);learnerEval.setLabelConverter(converter);
      int states = referenceGraph.getAcceptStateNumber();
      final Collection<List<Label>> testSet = PaperUAS.computeEvaluationSet(referenceGraph,states*3,PairQualityLearner.makeEven(states*referenceGraph.pathroutines.computeAlphabet().size()));
      
      DifferenceToReferenceDiff differenceStructural=DifferenceToReferenceDiff.estimationOfDifferenceDiffMeasure(referenceGraph, learntGraph, config, 1);
      DifferenceToReferenceLanguageBCR differenceBCRlearnt=DifferenceToReferenceLanguageBCR.estimationOfDifference(referenceGraph, learntGraph,testSet);
    
      final MarkovModel m= new MarkovModel(chunkLen,true,true,false);
      LearnerGraph pta=new LearnerGraph(config);
      for(List<Label> seq:sPlus)
        pta.paths.augmentPTA(seq,true,false,null);
      for(List<Label> seq:sMinus)
        pta.paths.augmentPTA(seq,false,false,null);
      pta.clearColours();
      new MarkovClassifier(m, pta).updateMarkov(false);// construct Markov chain
      // For Markov, we do not need to learn anything at all - our Markov matrix contains enough information to classify paths and hence compare it to the reference graph.
      ConfusionMatrix mat = DiffExperiments.classifyAgainstMarkov(testSet, referenceGraph, m);
      DifferenceToReferenceLanguageBCR differenceBCRMarkov = new DifferenceToReferenceLanguageBCR(mat);
      
      return new OtpErlangTuple(new OtpErlangObject[]{
          new OtpErlangDouble(differenceStructural.getValue()),new OtpErlangDouble(differenceBCRlearnt.getValue()),new OtpErlangDouble(differenceBCRMarkov.getValue())
      });
    }

View Full Code Here

TOP

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.