如何在 Java Weka API 中使用 class 不平衡技术 (SMOTE)？

Question

我正在尝试使用 Java Weka API 构建 class化模型。我的训练数据集有 class 不平衡问题。出于这个原因，我想使用像 SMOTE 这样的 class 不平衡技术来减少 class 不平衡问题。

源代码如下：

package classification;
import java.util.Random;
import weka.classifiers.Classifier;
import weka.classifiers.bayes.NaiveBayesMultinomial;
import weka.core.Instance;
import weka.core.Instances;
import weka.core.converters.ConverterUtils.DataSource;
import weka.filters.Filter;
import weka.filters.unsupervised.attribute.StringToWordVector;
public class questWhosebug {
public static void main(String agrs[]) throws Exception{
String fileRootPath = "../file.arff"; //Dataset
    Instances strdata = DataSource.read(fileRootPath); //Load Dataset
    StringToWordVector filter = new StringToWordVector(10000);
    filter.setInputFormat(strdata);
    String[] options = { "-W", "10000", "-L", "-M", "1",
            "-stemmer", "weka.core.stemmers.IteratedLovinsStemmer", 
            "-stopwords-handler", "weka.core.stopwords.Rainbow", 
            "-tokenizer", "weka.core.tokenizers.AlphabeticTokenizer" 
            };
    filter.setOptions(options);
    filter.setIDFTransform(true);
    Instances data = Filter.useFilter(strdata,filter); //Apply filter
    data.setClassIndex(0); //set class index        
    double recall=0.0;
    double precision=0.0;
    double fmeasure=0.0;
    double tp, fp, fn, tn;

    Classifier classifier = null;
    classifier = new NaiveBayesMultinomial(); //classifer

    int folds = 10;         
    Random random = new Random(1);
    data.randomize(random);
    data.stratify(folds);
    tp = fp = fn = tn = 0;
    for (int i = 0; i < folds; i++) {
       Instances trains = data.trainCV(folds, i,random); //training dataset
       Instances tests = data.testCV(folds, i); //testing dataset
        classifier.buildClassifier(trains);    //build classifier           
        for (int j = 0; j < tests.numInstances(); j++) {    
           Instance instance = tests.instance(j);
           double classValue = instance.classValue();                   
           double result = classifier.classifyInstance(instance);
            if (result == 0.0 && classValue == 0.0) {
                    tp++;
                } else if (result == 0.0 && classValue == 1.0) {
                    fp++;
                } else if (result == 1.0 && classValue == 0.0) {
                    fn++;
                } else if (result == 1.0 && classValue == 1.0) {
                    tn++;
                }
            }   
        }

        if (tn + fn > 0)
            precision = tn / (tn + fn);
        if (tn + fp > 0)
            recall = tn / (tn + fp);
        if (precision + recall > 0)
            fmeasure = 2 * precision * recall / (precision + recall);
        System.out.println("Precision: " + precision);
        System.out.println("Recall: " + recall);
        System.out.println("Fmeasure: " + fmeasure);

    }

}

我的代码在没有 class 不平衡技术的情况下运行良好。但是，我需要使用 class 不平衡技术来缓解 class 不平衡问题。但是，我不知道如何在 Java Weka API.

中使用它

Answer 1

您可以在代码中添加以下代码行：

weka.filters.supervised.instance.SMOTE


SMOTE smote=new SMOTE();
smote.setInputFormat(trains);       
Instances Trains_smote= Filter.useFilter(trains, smote);

您的代码如下。

package classification;
import java.util.Random;
import weka.classifiers.Classifier;
import weka.classifiers.bayes.NaiveBayesMultinomial;
import weka.core.Instance;
import weka.core.Instances;
import weka.core.converters.ConverterUtils.DataSource;
import weka.filters.Filter;
import weka.filters.unsupervised.attribute.StringToWordVector;
weka.filters.supervised.instance.SMOTE
public class questWhosebug {
public static void main(String agrs[]) throws Exception{
String fileRootPath = "../file.arff"; //Dataset
Instances strdata = DataSource.read(fileRootPath); //Load Dataset
StringToWordVector filter = new StringToWordVector(10000);
filter.setInputFormat(strdata);
String[] options = { "-W", "10000", "-L", "-M", "1",
        "-stemmer", "weka.core.stemmers.IteratedLovinsStemmer", 
        "-stopwords-handler", "weka.core.stopwords.Rainbow", 
        "-tokenizer", "weka.core.tokenizers.AlphabeticTokenizer" 
        };
filter.setOptions(options);
filter.setIDFTransform(true);
Instances data = Filter.useFilter(strdata,filter); //Apply filter
data.setClassIndex(0); //set class index        
double recall=0.0;
double precision=0.0;
double fmeasure=0.0;
double tp, fp, fn, tn;

Classifier classifier = null;
classifier = new NaiveBayesMultinomial(); //classifer

int folds = 10;         
Random random = new Random(1);
data.randomize(random);
data.stratify(folds);
tp = fp = fn = tn = 0;
for (int i = 0; i < folds; i++) {
   Instances trains = data.trainCV(folds, i,random); //training dataset
   Instances tests = data.testCV(folds, i); //testing dataset
   SMOTE smote=new SMOTE();
   smote.setInputFormat(trains);        
   Instances Trains_smote = Filter.useFilter(trains, smote);

    classifier.buildClassifier(Trains_smote);    //build classifier           
    for (int j = 0; j < tests.numInstances(); j++) {    
       Instance instance = tests.instance(j);
       double classValue = instance.classValue();                   
       double result = classifier.classifyInstance(instance);
        if (result == 0.0 && classValue == 0.0) {
                tp++;
            } else if (result == 0.0 && classValue == 1.0) {
                fp++;
            } else if (result == 1.0 && classValue == 0.0) {
                fn++;
            } else if (result == 1.0 && classValue == 1.0) {
                tn++;
            }
        }   
    }

    if (tn + fn > 0)
        precision = tn / (tn + fn);
    if (tn + fp > 0)
        recall = tn / (tn + fp);
    if (precision + recall > 0)
        fmeasure = 2 * precision * recall / (precision + recall);
    System.out.println("Precision: " + precision);
    System.out.println("Recall: " + recall);
    System.out.println("Fmeasure: " + fmeasure);

}

}

如何在 Java Weka API 中使用 class 不平衡技术 (SMOTE)？

How to use class Imbalance technique (SMOTE) with Java Weka API?

java

weka

text-classification