Question

import javax.print.attribute.DateTimeSyntax;
import weka.core.Instances;
import weka.core.converters.ConverterUtils.DataSource;
import weka.classifiers.bayes.NaiveBayes;
import weka.classifiers.trees.J48;
import weka.classifiers.functions.SMO;
public class Classification {
    public static void main(String[] args) throws Exception{
        //load dataset
        DataSource source = new DataSource("j:/weka/wekadataset/iris.arff");
        Instances dataset = source.getDataSet();
        // set the class to the last class attribute
        dataset.setClassIndex(dataset.numAttributes()-1);
        //create and build classifier
        J48 tree= new J48();
        tree.buildClassifier(dataset);
    }
}

//我在半监督学习中使用J48。在每次迭代之后，我必须访问标记的数据，以便我可以在下一次迭代中使用它。 我的问题在这里。如何访问每个步骤的标记数据？我应该更改J48源代码以便我可以看到它或者有办法执行此操作（在Java中为此目的，是否有方法？）

Answer 1

//write the modified dataset by j48 into the new file and
// write in the output
    BufferedWriter bw = new BufferedWriter(new OutputStreamWriter(fos));
    for (int i=0; i<k ; i++){
        Instance newInstance = instance01.instance(i);
        //write attributes 
        for(int j=0;j<newInstance.numAttributes()-1;j++){
            String AttribValue =Double.toString(newInstance.value(j));
            System.out.print(newInstance.value(j)+"-");
            bw.write(AttribValue+",");
        }
        //write the class
        double actualClass = tree.classifyInstance(newInstance);
        String actual =newInstance.classAttribute().value((int)actualClass);
        System.out.println("class="+actual);
        bw.write(actual);
        bw.newLine();
    }
    bw.close();

Answer 2

//now compare the source file and the modified dataset in the second
//file. write the differences in the third file and then rename this file
//to the second file
FileOutputStream fos1 = new FileOutputStream(new File
        ("j:/weka/wekadataset/new folder/iris-semi02.arff"));//third file
BufferedWriter bw1 = new BufferedWriter(new OutputStreamWriter(fos1));
BufferedReader br1 = new BufferedReader(new FileReader(
        "j:/weka/wekadataset/iris-semi.arff"));//
BufferedReader br2 = new BufferedReader(new FileReader(
        "j:/weka/wekadataset/new folder/iris-semi01.arff"));
String line1=br1.readLine();
String line2=br2.readLine();
while(!line1.startsWith("@DATA")){line1=br1.readLine();}
    line1=br1.readLine();   
    while(line1!=null && line2!=null ){
        if(!line1.equals(line2)){ 
            bw1.write(line2);   
            bw1.newLine();
        }
        line1=br1.readLine();
        line2=br2.readLine();
}
bw1.close();

如何访问标有j48的数据？

2 个答案: