audio event 實驗中的weka交叉代碼

import java.io.File;

import weka.classifiers.Classifier;
import weka.classifiers.Evaluation;
import weka.classifiers.trees.RandomForest;
import weka.core.Instance;
import weka.core.Instances;
import weka.core.converters.ArffLoader;

import java.util.*;

public class test {
	Vector<String> alllist=new Vector<String>();
	void getlist(){
		alllist.add("list/animal/cat.txt.arff");
		alllist.add("list/animal/dog.txt.arff");
		
		alllist.add("list/household/clock.txt.arff");
		alllist.add("list/household/door.txt.arff");
		
		alllist.add("list/musical+instruments/guitar.txt.arff");
		alllist.add("list/musical+instruments/trumpet.txt.arff");
		
		alllist.add("list/nature/thunder.txt.arff");
		alllist.add("list/nature/wind.txt.arff");
		
		alllist.add("list/office/coins.txt.arff");
		alllist.add("list/office/paper.txt.arff");
		
		alllist.add("list/people/applause.txt.arff");
		alllist.add("list/people/heartbeat.txt.arff");
		
		alllist.add("list/sports+and+recreation/bowling.txt.arff");
		alllist.add("list/sports+and+recreation/cards.txt.arff");
		
		alllist.add("list/vehicles/engine.txt.arff");
		alllist.add("list/vehicles/train.txt.arff");
	}
	Instances getallinstances(){
		this.getlist();
		Instances alldataset=null;
		for(int i=0;i<this.alllist.size();i++){
			String path=alllist.get(i);
			ArffLoader loader = new ArffLoader();
			Instances dataset=null;
			File file = new File(path);
			try{
				loader.setFile(file);
				System.out.println(path);
				dataset=loader.getDataSet();
				int numAttr=dataset.numAttributes();
				dataset.setClassIndex(numAttr-1);
				if(alldataset==null){
					alldataset=new Instances(dataset);
				}
				else{
					int num=dataset.numInstances();
					for(int j=0;j<num;j++){
						Instance instance=dataset.instance(j);
						alldataset.add(instance);
					}
				}
			}
			catch(Exception e){
				e.printStackTrace();
			}
		}
		return alldataset;
	}
	public static void main(String[] args) {
		test t=new test();
		 Instances dataset=null;
		 try{
			 dataset=t.getallinstances();
			 
			 int seed=5;
			 
			 int folds=5;
			 double errorRate=0;
			 double correRate=0;
			 Random rand=new Random(seed);
			 Instances randData=new Instances(dataset);
			 randData.randomize(rand);
			 for(int i=0;i<folds;i++){
				 
				 Instances traindata=randData.trainCV(folds, i);
				 Instances testdata=randData.testCV(folds, i);
				 System.out.println("train size is: "+traindata.numInstances());
				 //System.out.println("test size  is: "+testdata.numInstances());
				 Scanner scan=new Scanner(System.in);
				// scan.nextLine();
				 seed++;
				 /*
				 Classifier svm=new LibSVM();
				 svm.setOptions(weka.core.Utils.splitOptions("-S 0 -K 2 -D 3 -G 0.0 -R 0.0 -N 0.5 -M 40.0 -C 8.0 -E 0.004 -P 0.1 -seed 1"));
				 svm.buildClassifier(traindata);
				 Evaluation evaluation = new Evaluation(testdata);
				 evaluation.evaluateModel(svm,testdata );
				 errorRate+=evaluation.errorRate();
				 */
				 Classifier randforest=new RandomForest();
				 randforest.buildClassifier(traindata);
				 Evaluation evaluation = new Evaluation(testdata);
				 evaluation.evaluateModel(randforest,testdata );
				 //System.out.println("error rate is: "+evaluation.errorRate());
				 //System.out.println("precision rate is: "+evaluation.correct()/testdata.numInstances());
				 //scan.nextLine();
				 errorRate+=evaluation.errorRate();
				 correRate+=evaluation.correct()/testdata.numInstances();
				 double confusion[][]=evaluation.confusionMatrix();
				 System.out.println("confusion size is: "+confusion.length+" "+confusion[0].length);
				 for(int k=0;k<confusion.length;k++){
					 double []confu=confusion[k];
					 for(int kk=0;kk<confu.length;kk++){
						 System.out.print(confu[kk]+" ");
					 }
					 System.out.println();
				 }
				 scan.nextLine();
				 //confusion=confusion+confusion;
			 }
			 System.out.println("error rate is:"+errorRate/folds);
			 System.out.println("corre rate is:"+correRate/folds);
			 /*
			 int num=dataset.numAttributes();
			 for(int i=0;i<dataset.numInstances();i++){
				 Instance instance=dataset.instance(i);
				 System.out.println("class is: "+instance.toString(num-1));
			 }
			 System.out.println("number is: "+dataset.numInstances());
			 Classifier svm=new LibSVM();
			 svm.setOptions(weka.core.Utils.splitOptions("-S 0 -K 2 -D 3 -G 0.0 -R 0.0 -N 0.5 -M 40.0 -C 8.0 -E 0.004 -P 0.1 -seed 1"));
			 svm.buildClassifier(dataset);
			 Evaluation evaluation = new Evaluation(dataset);
			 evaluation.evaluateModel(svm, dataset);
			 System.out.println("error rate is: "+evaluation.errorRate());
			 
			 System.out.println(dataset.numInstances()+" "+evaluation.correct());
			 */
		 }
		 catch(Exception e){
			 e.printStackTrace();
		 }
	}
}