本文整理汇总了Java中weka.attributeSelection.InfoGainAttributeEval.buildEvaluator方法的典型用法代码示例。如果您正苦于以下问题：Java InfoGainAttributeEval.buildEvaluator方法的具体用法？Java InfoGainAttributeEval.buildEvaluator怎么用？Java InfoGainAttributeEval.buildEvaluator使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类weka.attributeSelection.InfoGainAttributeEval的用法示例。
在下文中一共展示了InfoGainAttributeEval.buildEvaluator方法的3个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Java代码示例。
示例1: InformationGain
import weka.attributeSelection.InfoGainAttributeEval; //导入方法依赖的package包/类
public InformationGain(Instances dataset) throws Exception{
data = dataset;
infoG = new InfoGainAttributeEval();
infoG.buildEvaluator(data);
filter = new AttributeSelection();
search = new Ranker();
search.setThreshold(0.0);
}
示例2: classify
import weka.attributeSelection.InfoGainAttributeEval; //导入方法依赖的package包/类
public void classify(String trainingFile,String testingFile) {
try {
initTrainingSet(trainingFile);
initTestingSet(testingFile);
// train NaiveBayes
J48 cModel = new J48();
cModel.buildClassifier(TrainingSet);
Instance current;
double pred=0;
for (int i = 0; i < TestingSet.numInstances(); i++) {
current=TestingSet.get(i);
if(featureVectorClassValues.get((int)pred).equalsIgnoreCase("Statement")||featureVectorClassValues.get((int)pred).equalsIgnoreCase("Backchannel Question")||featureVectorClassValues.get((int)pred).equalsIgnoreCase("Yes-No Question")||featureVectorClassValues.get((int)pred).equalsIgnoreCase("Open Question")){
current.setValue(featureVectorAttributes.get(0),featureVectorClassValues.get((int)pred));
System.out.println(pred+" : "+featureVectorClassValues.get((int)pred));
System.out.println(current.toString());
}
pred=cModel.classifyInstance(current);
}
// J48 cModel = new J48();
// cModel.setUnpruned(true);
// cModel.buildClassifier(TrainingSet);
Evaluation eTest = new Evaluation(TrainingSet);
eTest.evaluateModel(cModel, TestingSet);
//print out the results
System.out.println("=====================================================================");
System.out.println("Results for "+this.getClass().getSimpleName());
String strSummary = eTest.toSummaryString();
System.out.println(strSummary);
System.out.println("F-measure : "+eTest.weightedFMeasure());
System.out.println("precision : "+eTest.weightedPrecision());
System.out.println("recall : "+eTest.weightedRecall());
System.out.println("=====================================================================");
InfoGainAttributeEval infoGainAttributeEval = new InfoGainAttributeEval();
infoGainAttributeEval.buildEvaluator(TrainingSet);
for (int i = 0; i <featureVectorAttributes.size()-1; i++) {
double v = infoGainAttributeEval.evaluateAttribute(i);
System.out.print(featureVectorAttributes.get(i).name()+"\t\t");
System.out.println(v);
}
} catch (Exception e) {
e.printStackTrace();
}
}
示例3: classify
import weka.attributeSelection.InfoGainAttributeEval; //导入方法依赖的package包/类
public void classify(String trainingFile,String testingFile) {
try {
// initiateBagOfWords(trainingFile);
initTrainingSet(trainingFile);
// initiateBagOfWords(testingFile);
initTestingSet(testingFile);
StringToWordVector filter = new StringToWordVector();
int[] indices= new int[1];
indices[0]=6;
filter.setAttributeIndicesArray(indices);
filter.setInputFormat(TrainingSet);
filter.setWordsToKeep(6);
filter.setDoNotOperateOnPerClassBasis(false);
filter.setTFTransform(true);
filter.setOutputWordCounts(true);
TrainingSet = Filter.useFilter(TrainingSet, filter);
TestingSet = Filter.useFilter(TestingSet, filter);
Classifier cModel = new SimpleLogistic();
cModel.buildClassifier(TrainingSet);
weka.core.SerializationHelper.write(System.getProperty("user.dir")+"/Classification/src/datafiles/cls.model",cModel);
weka.core.SerializationHelper.write(System.getProperty("user.dir")+"/Classification/src/datafiles/testingSet.model",TestingSet);
Evaluation eTest = new Evaluation(TrainingSet);
eTest.evaluateModel(cModel, TestingSet);
//print out the results
System.out.println("=====================================================================");
System.out.println("Results for "+this.getClass().getSimpleName());
String strSummary = eTest.toSummaryString();
System.out.println(strSummary);
InfoGainAttributeEval infoGainAttributeEval = new InfoGainAttributeEval();
infoGainAttributeEval.buildEvaluator(TrainingSet);
for (int i = 0; i <featureVectorAttributes.size()-1; i++) {
double v = infoGainAttributeEval.evaluateAttribute(i);
System.out.print(i+" "+featureVectorAttributes.get(i).name()+"\t\t");
System.out.println(v);
}
System.out.println("=====================================================================");
System.out.println("recall : "+eTest.weightedRecall());
System.out.println("precision : "+eTest.weightedPrecision());
System.out.println("F-measure : "+eTest.weightedFMeasure());
System.out.println("================= Rounded Values =========================");
System.out.println("recall : "+Math.round(eTest.weightedRecall() * 100.0) / 100.0);
System.out.println("precision : "+Math.round(eTest.weightedPrecision() * 100.0) / 100.0);
System.out.println("F-measure : "+Math.round(eTest.weightedFMeasure() * 100.0) / 100.0);
System.out.println("=====================================================================");
printErrors(cModel);
} catch (Exception e) {
e.printStackTrace();
}
}