Examples of weka.experiment.SplitEvaluator

weka.experiment.SplitEvaluator
Interface to objects able to generate a fixed set of results for a particular split of a dataset. The set of results should contain fields related to any settings of the SplitEvaluator (not including the dataset name. For example, one field for the classifier used to get the results, another for the classifier options, etc).
Possible implementations of SplitEvaluator:
- StdClassification results
- StdRegression results
@author Len Trigg (trigg@cs.waikato.ac.nz) @version $Revision: 1.7 $

      m_NumberOfRepetitionsTField.setEnabled(true);
      m_Exp.setRunLower(1);
      m_Exp.setRunUpper(m_numRepetitions);
    }


    SplitEvaluator se = null;
    Classifier sec = null;
    if (m_ExpClassificationRBut.isSelected()) {
      se = new ClassifierSplitEvaluator();
      sec = ((ClassifierSplitEvaluator)se).getClassifier();
    } else {
      se = new RegressionSplitEvaluator();
      sec = ((RegressionSplitEvaluator)se).getClassifier();
    }
    
    // build new ResultProducer
    if (m_ExperimentTypeCBox.getSelectedItem() == TYPE_CROSSVALIDATION_TEXT) {
      CrossValidationResultProducer cvrp = new CrossValidationResultProducer();
      cvrp.setNumFolds(m_numFolds);
      cvrp.setSplitEvaluator(se);
      
      PropertyNode[] propertyPath = new PropertyNode[2];
      try {
  propertyPath[0] = new PropertyNode(se, new PropertyDescriptor("splitEvaluator",
                      CrossValidationResultProducer.class),
             CrossValidationResultProducer.class);
  propertyPath[1] = new PropertyNode(sec, new PropertyDescriptor("classifier",
                       se.getClass()),
             se.getClass());
      } catch (IntrospectionException e) {
  e.printStackTrace();
      }
      
      m_Exp.setResultProducer(cvrp);
      m_Exp.setPropertyPath(propertyPath);


    } else {
      RandomSplitResultProducer rsrp = new RandomSplitResultProducer();
      rsrp.setRandomizeData(m_ExperimentTypeCBox.getSelectedItem() == TYPE_RANDOMSPLIT_TEXT);
      rsrp.setTrainPercent(m_trainPercent);
      rsrp.setSplitEvaluator(se);


      PropertyNode[] propertyPath = new PropertyNode[2];
      try {
  propertyPath[0] = new PropertyNode(se, new PropertyDescriptor("splitEvaluator",
                      RandomSplitResultProducer.class),
             RandomSplitResultProducer.class);
  propertyPath[1] = new PropertyNode(sec, new PropertyDescriptor("classifier",
                       se.getClass()),
             se.getClass());
      } catch (IntrospectionException e) {
  e.printStackTrace();
      }


      m_Exp.setResultProducer(rsrp);

View Full Code Here

      m_NumberOfRepetitionsTField.setEnabled(true);
      m_Exp.setRunLower(1);
      m_Exp.setRunUpper(m_numRepetitions);
    }


    SplitEvaluator se = null;
    Classifier sec = null;
    if (m_ExpClassificationRBut.isSelected()) {
      se = new ClassifierSplitEvaluator();
      sec = ((ClassifierSplitEvaluator)se).getClassifier();
    } else {
      se = new RegressionSplitEvaluator();
      sec = ((RegressionSplitEvaluator)se).getClassifier();
    }
    
    // build new ResultProducer
    if (m_ExperimentTypeCBox.getSelectedItem() == TYPE_CROSSVALIDATION_TEXT) {
      CrossValidationResultProducer cvrp = new CrossValidationResultProducer();
      cvrp.setNumFolds(m_numFolds);
      cvrp.setSplitEvaluator(se);
      
      PropertyNode[] propertyPath = new PropertyNode[2];
      try {
  propertyPath[0] = new PropertyNode(se, new PropertyDescriptor("splitEvaluator",
                      CrossValidationResultProducer.class),
             CrossValidationResultProducer.class);
  propertyPath[1] = new PropertyNode(sec, new PropertyDescriptor("classifier",
                       se.getClass()),
             se.getClass());
      } catch (IntrospectionException e) {
  e.printStackTrace();
      }
      
      m_Exp.setResultProducer(cvrp);
      m_Exp.setPropertyPath(propertyPath);


    } else {
      RandomSplitResultProducer rsrp = new RandomSplitResultProducer();
      rsrp.setRandomizeData(m_ExperimentTypeCBox.getSelectedItem() == TYPE_RANDOMSPLIT_TEXT);
      rsrp.setTrainPercent(m_trainPercent);
      rsrp.setSplitEvaluator(se);


      PropertyNode[] propertyPath = new PropertyNode[2];
      try {
  propertyPath[0] = new PropertyNode(se, new PropertyDescriptor("splitEvaluator",
                      RandomSplitResultProducer.class),
             RandomSplitResultProducer.class);
  propertyPath[1] = new PropertyNode(sec, new PropertyDescriptor("classifier",
                       se.getClass()),
             se.getClass());
      } catch (IntrospectionException e) {
  e.printStackTrace();
      }


      m_Exp.setResultProducer(rsrp);

View Full Code Here

      m_NumberOfRepetitionsTField.setEnabled(true);
      m_Exp.setRunLower(1);
      m_Exp.setRunUpper(m_numRepetitions);
    }


    SplitEvaluator se = null;
    Classifier sec = null;
    if (m_ExpClassificationRBut.isSelected()) {
      se = new ClassifierSplitEvaluator();
      sec = ((ClassifierSplitEvaluator)se).getClassifier();
    } else {
      se = new RegressionSplitEvaluator();
      sec = ((RegressionSplitEvaluator)se).getClassifier();
    }
    
    // build new ResultProducer
    if (m_ExperimentTypeCBox.getSelectedItem() == TYPE_CROSSVALIDATION_TEXT) {
      CrossValidationResultProducer cvrp = new CrossValidationResultProducer();
      cvrp.setNumFolds(m_numFolds);
      cvrp.setSplitEvaluator(se);
      
      PropertyNode[] propertyPath = new PropertyNode[2];
      try {
  propertyPath[0] = new PropertyNode(se, new PropertyDescriptor("splitEvaluator",
                      CrossValidationResultProducer.class),
             CrossValidationResultProducer.class);
  propertyPath[1] = new PropertyNode(sec, new PropertyDescriptor("classifier",
                       se.getClass()),
             se.getClass());
      } catch (IntrospectionException e) {
  e.printStackTrace();
      }
      
      m_Exp.setResultProducer(cvrp);
      m_Exp.setPropertyPath(propertyPath);


    } else {
      RandomSplitResultProducer rsrp = new RandomSplitResultProducer();
      rsrp.setRandomizeData(m_ExperimentTypeCBox.getSelectedItem() == TYPE_RANDOMSPLIT_TEXT);
      rsrp.setTrainPercent(m_trainPercent);
      rsrp.setSplitEvaluator(se);


      PropertyNode[] propertyPath = new PropertyNode[2];
      try {
  propertyPath[0] = new PropertyNode(se, new PropertyDescriptor("splitEvaluator",
                      RandomSplitResultProducer.class),
             RandomSplitResultProducer.class);
  propertyPath[1] = new PropertyNode(sec, new PropertyDescriptor("classifier",
                       se.getClass()),
             se.getClass());
      } catch (IntrospectionException e) {
  e.printStackTrace();
      }


      m_Exp.setResultProducer(rsrp);

View Full Code Here

TOP

Related Classes of weka.experiment.SplitEvaluator

weka.gui.experiment.SimpleSetupPanel

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.