アットウィキロゴ

kusai

package ma;
 
import java.io.*;
import java.util.*;
import java.util.regex.*;
 
import cc.mallet.pipe.*;
import cc.mallet.pipe.iterator.*;
import cc.mallet.types.*;
 
public class pro {
 
    Pipe pipe;
 
    public static void main (String[] args)  {
 
        pro test = new pro();
      }
 
 
    public pro() {
        pipe = play();
    }
 
    public Pipe play() {
 
        ArrayList p2 = new ArrayList();
 
 
        p2.add(new Input2CharSequence("UTF-8"));
 
        Pattern tk =Pattern.compile("[\\p{L}\\p{N}_]+");
 
        p2.add(new CharSequence2TokenSequence(tk));
 
      p2.add(new TokenSequenceLowercase());
 
 
        p2.add(new TokenSequenceRemoveStopwords(false, false));
 
 
        p2.add(new TokenSequence2FeatureSequence());
 
 
        p2.add(new Target2Label());
 
 
        p2.add(new FeatureSequence2FeatureVector());
 
 
      p2.add(new PrintInputAndTarget());
 
        return new SerialPipes(p2);
    }
 
 
    }
 
最終更新:2016年05月03日 05:12