Question

我一直在尝试使用Standford Weka的MISVM。但我一直得到错误

weka.classifiers.mi.MISVM: Cannot handle numeric attributes!

我的训练集文件（file.arff）

@relation 'SpamDetectionTrainingSet'
@attribute 'checkin' real
@attribute 'tips' real
@attribute 'badges' real
@attribute 'friends' real
@attribute 'mayors' real
@attribute 'lists' real
@attribute 'target' {'spam','nonspam'}
@data
1165,36,1,1,4,47,nonspam
8521,492,620,7,74,144,nonspam
5408,189,306,0,27,121,nonspam
0,43,319,0,0,0,spam

我已经看到了之前问题的解决方案。我从他们身上观察到的一个要点是保持target attribute名义上的。我的目标属性只是名义上的

我的源代码

import weka.core.Instances;
import weka.core.Utils;
import weka.core.converters.ConverterUtils.DataSink;
import weka.core.converters.ConverterUtils.DataSource;
public class SVM 
{
    /* Read the Data in ARFF format */
    DataSource train,unknown,test;
    Instances trainData,unknownData;
    public SVM()
    {
        try {
            /* Loading the Data */
            train       = new DataSource("/home/hduser/file.arff");
            trainData   = train.getDataSet();
            trainData.setClassIndex(trainData.numAttributes()-1);

            /* Set Option for SVM Algorithm */
            String[] opt = Utils.splitOptions("-N 1 -D");

            /*Training */
            weka.classifiers.mi.MISVM rf = new weka.classifiers.mi.MISVM();

            rf.setOptions(opt);
            System.out.println(rf.getFilterType());

            rf.buildClassifier(trainData);

            /* Classify Users */
            unknown     = new DataSource("/home/hduser/fileTest1.arff");
            unknownData = unknown.getDataSet();
            unknownData.setClassIndex(unknownData.numAttributes()-1);

            Instances copy = new Instances(unknownData);

            for(int i=0;i<unknownData.numInstances();i++)
            {
                System.out.println("Hello ");
                double value = rf.classifyInstance(unknownData.instance(i));
                copy.instance(i).setClassValue(value);
            }

            DataSink.write("/home/hduser/output1.arff", copy);
        } 
        catch (Exception e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
        }
    }
    public static void main(String[] args)
    {
        SVM r = new SVM();
    }
}

完整的错误消息：

weka.core.UnsupportedAttributeTypeException: weka.classifiers.mi.MISVM: Cannot handle numeric attributes!
    at weka.core.Capabilities.test(Capabilities.java:954)
    at weka.core.Capabilities.test(Capabilities.java:869)
    at weka.core.Capabilities.test(Capabilities.java:1085)
    at weka.core.Capabilities.test(Capabilities.java:1023)
    at weka.core.Capabilities.testWithFail(Capabilities.java:1302)
    at weka.classifiers.mi.MISVM.buildClassifier(MISVM.java:607)
    at SVM.<init>(SVM.java:28)
    at SVM.main(SVM.java:53)

Answer 1

问题是file.arff：

@data
1165,36,1,1,4,47,nonspam
8521,492,620,7,74,144,nonspam
5408,189,306,0,27,121,nonspam
0,43,319,0,0,0,spam

没有数值。

更改数据集，或更改分类器

无法处理数字属性weka svm

1 个答案: