Java weka分类返回分类名称classname

使用eclipse 、Java 、weka

---

1、在weka的安装目录下找到weka.jar和weka-src.jar两个jar包，如下图所示：

2、在eclipse中新建Java项目，然后右键build-path，在Libraries标签页里点击添加外部jar吧，然后将上一步中找到的weka.jar和weka-src.jar添加进去，然后点击OK，如下图所示：

3、在新建的Java项目中新建包wekaTest和类J48Test，将分词后的中文文本添加筛选器StringToWordVector（对文本数据进行预处理），然后使用weka自带的分类算法进行分类，并返回每个实例对应的分类名称，具体实现代码如下：

package wekaTest;

import java.io.File;
import weka.classifiers.Evaluation;
import weka.classifiers.bayes.NaiveBayes;
import weka.classifiers.trees.J48;
import weka.classifiers.trees.RandomForest;
import weka.core.Instance;
import weka.core.Instances;
import weka.core.converters.ArffLoader;
import weka.filters.Filter;
import weka.filters.unsupervised.attribute.StringToWordVector;

public class J48Test {
    public static void main(String[] args) throws Exception {
        Instances ins = null;

        try {
               File file = new File("E:\\data.arff");
               ArffLoader loader = new ArffLoader();
               loader.setFile(file);
               ins = loader.getDataSet();              
               ins.setClassIndex(ins.numAttributes()-1);
               StringToWordVector filter = new StringToWordVector();

               filter.setIDFTransform(true);
               filter.setTFTransform(true);
               filter.setInputFormat(ins);
               Instances newIns = Filter.useFilter(ins, filter);
               //System.out.println(newIns.toString());
               RandomForest RFmodel = new RandomForest();
               J48 Jmodel = new J48();
               NaiveBayes NBmodel = new NaiveBayes();

               RFmodel.buildClassifier(newIns);
               Jmodel.buildClassifier(newIns);
               NBmodel.buildClassifier(newIns);

               Instance testInst;
               Evaluation testingEvaluationRF = new Evaluation(newIns);
               Evaluation testingEvaluationJ48 = new Evaluation(newIns);
               Evaluation testingEvaluationNB = new Evaluation(newIns);

               int length = newIns.numInstances();
               for(int i = 0; i < length ; i++){
                  testInst = newIns.instance(i);                         //System.out.println(newIns.classAttribute().value((int) RFmodel.classifyInstance(newIns.instance(i))));
//输出分类的类名                 //System.out.println(testInst.classAttribute().value((int) RFmodel.classifyInstance(testInst)));
                  testingEvaluationRF.evaluateModelOnceAndRecordPrediction(RFmodel, testInst);
                  testingEvaluationJ48.evaluateModelOnceAndRecordPrediction(Jmodel, testInst);
                  testingEvaluationNB.evaluateModelOnceAndRecordPrediction(NBmodel, testInst);
               }
               System.out.println("RandomForest的正确率："+(1-testingEvaluationRF.errorRate()));
               System.out.println("J48的正确率："+(1-testingEvaluationJ48.errorRate()));
               System.out.println("NaiveBayes的正确率："+(1-testingEvaluationNB.errorRate()));

//             System.out.println("RandomForest："+testingEvaluationRF.toSummaryString());//输出总结信息
//             System.out.println("RandomForest："+testingEvaluationRF.toClassDetailsString());//输出分类详细信息
//             System.out.println("RandomForest："+testingEvaluationRF.toMatrixString());//输出分类的混淆矩阵
          } catch (Exception e) {
               e.printStackTrace();
          }
    }
}

然后运行查看结果