java调用weka里面的分类器做回归

这里是M5P分类器，源文档有很大帮助

package rf;
import java.io.BufferedReader;
import java.io.FileReader;
import java.io.PrintStream;

import weka.classifiers.rules.M5Rules;
import weka.classifiers.trees.DecisionStump;
import weka.classifiers.trees.M5P;
import weka.classifiers.trees.RandomForest;
import weka.core.converters.ConverterUtils.DataSource;
import weka.core.Instances;

 
public class randomf {
public void runrf(int set) throws Exception{
	DataSource dataSource=new DataSource("svm/train_"+set+".arff");
	Instances data=dataSource.getDataSet();
	data.setClassIndex(data.numAttributes()-1);
	RandomForest rForest=new RandomForest();
	rForest.buildClassifier(data);//train
	Instances predictInstances=new Instances(new BufferedReader(new FileReader("svm/predict_"+set+".arff")));
	predictInstances.setClassIndex(predictInstances.numAttributes()-1);
	int sum=predictInstances.numAttributes();
	for (int i = 0; i < sum; i++) {
		double real=predictInstances.instance(i).classValue();
		double dis=rForest.classifyInstance(predictInstances.instance(i));
		System.out.println(real+" "+dis);
		
	}
	System.out.println("end");
}

//上面是随机怎林的分类，不能用于回归
public void m5p(int set,int min) throws Exception{
	PrintStream printStream=new PrintStream("set"+set+".predict");//文件输出
	DataSource dataSource=new DataSource("svm/train_"+set+".arff");//读取训练文件
	Instances data=dataSource.getDataSet();
	data.setClassIndex(data.numAttributes()-1);//最后一个是label
	//RandomForest rForest=new RandomForest();
	M5P m5p=new M5P();

	String[] opString=new String[3];
	opString[0]="-R";
	opString[1]="-M";
	opString[2]=Integer.toString(min);
	m5p.setOptions(opString);//设置参数

	m5p.buildClassifier(data);//train
	Instances predictInstances=new Instances(new BufferedReader(new FileReader("svm/predict_"+set+".arff")));//predict
	predictInstances.setClassIndex(predictInstances.numAttributes()-1);
	int sum=predictInstances.numInstances();
	System.out.println(sum);
	for (int i = 0; i < sum; i++) {
		//double real=predictInstances.instance(i).classValue();
		double dis=m5p.classifyInstance(predictInstances.instance(i));
		printStream.println(dis);
		
	}
	printStream.close();
	System.out.println("set "+set+"finished");	
}
}