中文新闻文本分类

 

下载搜狗新闻语料库

 

 

1.　　下载ictclas4j

3.　　导入外部包commons-lang-2.0.jar和ectclas4j.jar包

2)   按照上述第1,2，3的步骤配置好ictclas4j

3)   在weka文件目录中新建destFile目录（用于存放分词和去停用词后的结果）和srcFile目录（用于存放需要分词的文本文件和停用词表）

import java.io.BufferedReader;
import java.io.BufferedWriter;
import java.io.File;
import java.io.FileInputStream;
import java.io.FileNotFoundException;
import java.io.FileOutputStream;
import java.io.InputStreamReader;
import java.io.OutputStreamWriter;
import java.util.*;

import org.ictclas4j.bean.SegResult;  
import org.ictclas4j.segment.SegTag;
//import ICTCLAS.I3S.AC.ICTCLAS50;

public class FileExcludeStopWord {
    //停用词词表
    public static final String stopWordTable = "." + File.separator + "srcFile" + File.separator + "StopWordTable.txt";

    public static void main(String[] args) {

        //源文件和目的文件
        String srcFile = "." + File.separator + "srcFile" + File.separator + "酒店评论.txt";
        String destFile = "." + File.separator + "destFile" + File.separator + "酒店评论.txt";
        new FileExcludeStopWord().fileExcludeStopWord(srcFile, destFile);
    }
    
    public void fileExcludeStopWord(String srcFile,String destFile){
        try {
            //读取原文件和停用词表
            BufferedReader srcFileBr = new BufferedReader(new InputStreamReader(new FileInputStream(new File(srcFile))));
            BufferedReader StopWordFileBr = new BufferedReader(new InputStreamReader(new FileInputStream(new File(stopWordTable))));
            
            //将去除停用词的文本信息存入输出文件
            BufferedWriter destFileBw = new BufferedWriter(new OutputStreamWriter(new FileOutputStream(new File(destFile))));
            
            //用来存放停用词的集合
            Set<String> stopWordSet = new HashSet<String>();
            
            //初始化停用词集
            String stopWord = null;
            for(; (stopWord = StopWordFileBr.readLine()) != null;){
                stopWordSet.add(stopWord);        
            }            
            
            String paragraph = null;
            for(; (paragraph = srcFileBr.readLine()) != null;){
                //对读入的文本进行分词
                SegTag segTag = new SegTag(1);// 分词路径的数目          
                SegResult segResult = segTag.split(paragraph);
                String spiltResultStr = segResult.getFinalResult();    
                //得到分词后的词汇数组，以便后续比较
                String[] resultArray = spiltResultStr.split(" ");
                                
                //过滤停用词            
                for(int i = 0; i< resultArray.length; i++){
                    //System.out.println(resultArray[i]);
                    if(stopWordSet.contains(resultArray[i])){
                        resultArray[i] = null;
                    }
                    //System.out.println(resultArray[i]);    
                }
                
                //把过滤后的字符串数组存入到一个字符串中
                StringBuffer finalStr = new StringBuffer();
                for(int i = 0; i< resultArray.length; i++){
                    if(resultArray[i] != null){
                        finalStr = finalStr.append(resultArray[i]).append(" ");
                    }
                }
                
                //将过滤后的文本信息写入到指定文件中
                destFileBw.write(finalStr.toString());
                destFileBw.newLine();
                //输出最后的去停用词之后的结果
                System.out.println(finalStr);
            }
            
            //关闭输入流
            destFileBw.close();
            StopWordFileBr.close();
            srcFileBr.close();            
            
        } catch (FileNotFoundException e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
        } catch(Exception e){
            e.printStackTrace();
        }
    }
}

 java weka.core.converters.TextDirectoryLoader -dir training_1.0 > my_arff_save_path/data.arff

 

java weka.filters.unsupervised.attribute.StringToWordVector -I -C -i my_arff_save_path/data.arff -o my_arff_save_path/data_vsm.arff -c last 

java weka.filters.supervised.attribute.Discretize -i my_arff_save_path/data_vsm.arff -o my_arff_save_path/data_D_vsm.arff -c first

得到data_D_vsm.arff

 

 

 

测试

 

public class weka {

             BufferedWriter destFileBw = new BufferedWriter( new OutputStreamWriter(new FileOutputStream(newFile("E://data.txt" ))));

             File inputFile = new File( "E://data_D_vsm.arff" );//训练语料文件

             inputFile = new File( "E://data_D_vsm.arff" );//测试语料文件

             double sum = instancesTest .numInstances(), //测试语料实例数

             m_classifier.buildClassifier( instancesTrain ); //训练         

             for( int   i = 0; i < sum ; i ++) //测试分类结果

               double predicted = m_classifier.classifyInstance( instancesTest .instance( i));

                 System. out.println( "预测某条记录的分类id：" + predicted + ", 分类值："

                         + instancesTest.classAttribute().value(( int) predicted));

                 destFileBw.write( "预测某条记录的分类id：" + predicted + ", 分类值："

                         + instancesTest.classAttribute().value(( int) predicted));

                 System. out.println( "测试文件的分类值： " + instancesTest .instance(i ).classValue() + ", 记录："

                         + instancesTest.instance( i ));

                 destFileBw.write( "测试文件的分类值： " + instancesTest .instance(i).classValue() + ", 记录："

                         + instancesTest.instance( i ));

                 System.out .println("--------------------------------------------------------------" );

                 destFileBw.write( "--------------------------------------------------------------" );

                 if( m_classifier .classifyInstance(instancesTest .instance( i ))== instancesTest .instance(i).classValue())//如果预测值和答案值相等（测试语料中的分类列提供的须为正确答案，结果才有意义）

             System. out.println( "J48 classification precision:" +(right /sum ));

             destFileBw.write( "J48 classification precision:" +(right /sum ));