Twitter：weka.classifiers.bayes.NaiveBayesMultinomialUpdateable：无法处理字符串属性

Question

几天前我已经开始在 weka 上工作了。 （即我也是 weka 和 java 的新手）

我正在使用 weka 为推文寻找评估措施。运行程序时出现以下错误：

weka.classifiers.bayes.NaiveBayesMultinomialUpdateable：无法处理字符串属性！ weka.core.UnsupportedAttributeTypeException：weka.classifiers.bayes.NaiveBayesMultinomialUpdateable：无法处理字符串属性！ 在 weka.core.Capabilities.test（未知来源） 在 weka.core.Capabilities.test（未知来源） 在 weka.core.Capabilities.test（未知来源） 在 weka.core.Capabilities.test（未知来源） 在 weka.core.Capabilities.testWithFail（未知来源） 在 weka.classifiers.bayes.NaiveBayesMultinomialUpdateable.buildClassifier（未知来源） 在 com.ConnectGlobe.TextDirectoryToArff.main(TextDirectoryToArff.java:83)

计划：

package com.ConnectGlobe;
import java.io.*;
import weka.classifiers.Classifier;
import weka.classifiers.Evaluation;
import weka.classifiers.bayes.NaiveBayesMultinomialUpdateable;
import weka.core.*;
/**
 *
 * @author sv
 */
public class TextDirectoryToArff {

  public Instances createDataset(String directoryPath) throws Exception {

    FastVector atts = new FastVector(2);
    atts.addElement(new Attribute("filename", (FastVector) null));
    atts.addElement(new Attribute("contents", (FastVector) null));
    Instances data = new Instances("text_files_in_" + directoryPath, atts, 0);

    File dir = new File(directoryPath);
    String[] files = dir.list();
    for (int i = 0; i < files.length; i++) {
      if (files[i].endsWith(".txt")) {
    try {
      double[] newInst = new double[2];
      newInst[0] = (double)data.attribute(0).addStringValue(files[i]);
      File txt = new File(directoryPath + File.separator + files[i]);
      InputStreamReader is;
      is = new InputStreamReader(new FileInputStream(txt));
      StringBuffer txtStr = new StringBuffer();
      int c;
      while ((c = is.read()) != -1) {
        txtStr.append((char)c);
      }
      newInst[1] = (double)data.attribute(1).addStringValue(txtStr.toString());
      data.add(new Instance(1.0, newInst));
    } catch (Exception e) {
      //System.err.println("failed to convert file: " + directoryPath + File.separator + files[i]);
    }
      }
    }
    return data;
  }

  public static void main(String[] args) {

     TextDirectoryToArff tdta1 = new TextDirectoryToArff();
     TextDirectoryToArff tdta2 = new TextDirectoryToArff();
      try {
    Instances dataset1 = tdta1.createDataset("C:\\1"); // .txt file will be loaded
    dataset1.setClassIndex(dataset1.numAttributes() - 1 );
    Instances dataset2 = tdta2.createDataset("C:\\2");
    dataset2.setClassIndex(dataset1.numAttributes() - 1);

    System.out.println(dataset1);
    System.out.println(dataset2);

    double precision = 0, recall=0,fmeasure=0,error=0;

    int size1 = dataset1.numInstances() / 10;
    int size2 = dataset2.numInstances() / 10;

    int begin = 0;
    int end = size1 - 1 ;

    for (int i=1 ; i<=10;i++)
    {
        System.out.println("iteration :" + 1);

        Instances training = new Instances(dataset1);
        Instances testing = new Instances(dataset1, begin , (end - begin));

        for (int j=0;j < (end - begin); j++)
            training.delete(begin);

        Classifier tree = (Classifier)new NaiveBayesMultinomialUpdateable();


    Instances filteredInstaces = training;
    StringToNominal nominal ;


    for(int a=0;a<training.numAttributes()-1;a++)
    {
        if(training.attribute(a).isString())
        {
            nominal = new StringToNominal();
            nominal.setInputFormat(filteredInstaces);
            training = Filter.useFilter(training, nominal);
        }
    }

        tree.buildClassifier(filteredInstaces);

        Evaluation eval = new Evaluation(testing);

        eval.evaluateModel(tree, testing);
        System.out.println("Precision:" + eval.precision(1));
        System.out.println("Recall:" + eval.recall(1));
        System.out.println("Fmeasure:" + eval.fMeasure(1));
        System.out.println("Error:" + eval.errorRate());

        precision += eval.precision(1);
        recall += eval.recall(1);
        fmeasure += eval.fMeasure(1);
        error += eval.errorRate();

        //update 
        begin = end + 1;
        end+= size1;

        if(i==(9))
        {
            end = dataset1.numInstances();
        }

        System.out.println("Precision:" + precision/10.0);
        System.out.println("Recall:" + recall/10.0);
        System.out.println("Fmeasure:" + fmeasure/10.0);
        System.out.println("Error:" + error/0.0);
    }


   // Classifier cls = new NaiveBayesMultinomialUpdateable();
// cls.buildClassifier(dataset1);
 // evaluate classifier and print some statistics
// Evaluation eval = new Evaluation(dataset1);
 //eval.evaluateModel(cls, dataset2);
 //eval.crossValidateModel(cls,dataset1,10, dataset2.getRandomNumberGenerator(1));   
 //System.out.println(eval.toSummaryString("\nResults\n======\n", false));

      } catch (Exception e) {
    System.err.println(e.getMessage());
    e.printStackTrace();
      }
  }
}

Answer 1

Weka 中的朴素贝叶斯分类器不接受字符串属性，您需要先将字符串属性转换为名义属性。 Weka 有一个用于此目的的过滤器 StringToNominal，假设所有可能的字符串值都表示在您的第一个数据集（训练集）中，它应该可以工作。