Package org.fnlp.nlp.cn.tag

Examples of org.fnlp.nlp.cn.tag.CWSTagger

79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
   * @param args
   * @throws IOException
   * @throws 
   */
  public static void main(String[] args) throws Exception {
    CWSTagger tag = new CWSTagger("./models/seg.m"); 

    RLSeg rlseg = new RLSeg(tag,"./tmpdata/FNLPDATA/all.dict");

    tag.setDictionary(tempdict);

    System.out.println("\n处理文件：");
    String s4 = tag.tagFile("../example-data/data-tag.txt");
    System.out.println(s4);
    String[] toks = s4.split("\\s+");
    int newset = rlseg.update(toks);
    rlseg.close();

    tag.setDictionary(tempdict);

    System.out.println("\n处理文件：");
    String s = tag.tagFile("../example-data/data-tag.txt");
    System.out.println(s);


  }
View Full Code Here
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
   * @param args
   * @throws IOException
   * @throws 
   */
  public static void main(String[] args) throws Exception {
    CWSTagger tag = new CWSTagger("./models/seg.m");
    Dictionary dict=new Dictionary();
    dict.addFile("./models/dict.txt");
    tag.setDictionary(dict);
    ArrayList<String> str = MyCollection.loadList("./testcase/test case seg.txt",null);
    for(String s:str){     
      String t = tag.tag(s);
//      t = tag.tag(t);
      System.out.println(t);
    }
    tag.setEnFilter(false);
    for(String s:str){
      String t = tag.tag(s);
      System.out.println(t);
    }
   
    String t = tag.tagFile("data/FNLPDATA/seg/bad case.txt");
    System.out.println(t);
   
  }
View Full Code Here
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
   * @param args
   * @throws IOException
   * @throws 
   */
  public static void main(String[] args) throws Exception {
    CWSTagger tag = new CWSTagger("./models/seg.m");
    ArrayList<String> str = MyCollection.loadList("./testcase/seg.txt",null);
    for(String s:str){     
      String t = tag.tag(s);
//      t = tag.tag(t);
      System.out.println(t);
    }
    tag.setEnFilter(false);
    for(String s:str){
      String t = tag.tag(s);
      System.out.println(t);
    }
   
    ArrayList<String> str1 = MyCollection.loadList("data/FNLPDATA/seg/bad case.txt",null);
    for(String s:str1){   
      s = s.trim();
      String s1 = s.replaceAll(" ", "");
      String t = tag.tag(s1);
      System.out.println("处理： "+t);
      if(!t.equals(s))
        System.err.println("正确： "+s); 
    }
   
View Full Code Here
45
46
47
48
49
50
51
52
53
54
55
    for(m=1;m<=4;m++){
      System.out.println(m);
      MemoryStatic.start();
      switch(m){
      case 1:
        stag = new CWSTagger("../models/seg.m");
        System.out.println("分词");
        break;
      case 2:
        ttag = new POSTagger("../models/seg.m","../models/pos.m");
        System.out.println("词性标注");
View Full Code Here
58
59
60
61
62
63
64
65
66
67
68
  /**
   * @param args
   * @throws Exception
   */
  public static void main(String[] args) throws Exception {
    seg = new CWSTagger("./models/seg.m");
    cl = seg.getClassifier();
    int ysize = cl.getAlphabetFactory().getLabelSize();
    LinearViterbi vit = (LinearViterbi) cl.getInferencer();
    System.out.println(cl.getAlphabetFactory().getFeatureSize());
    HigherOrderViterbi inferencer = new HigherOrderViterbi(vit.getTemplets(), ysize);
View Full Code Here
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
    out.close();

  }
 
  private static void seg() throws IOException, LoadModelException {
    CWSTagger seg = new CWSTagger("../models/seg.m");
   
    BufferedReader in = new BufferedReader(new InputStreamReader(
        new FileInputStream(simpfile ), "utf8"));

    BufferedWriter out = new BufferedWriter(new OutputStreamWriter(new FileOutputStream(
        segfile), "utf8"));

    String line = null; 
    int count=0;
    while ((line = in.readLine()) != null) {
      if(line.length()==0){
       
      }else if(line.startsWith("<doc")){
        count++;       
      }else if(line.startsWith("</doc>")){
        count--;
      }else{
        line = seg.tag(line);     
      }
      out.append(line);
      out.append("\n");
    }
    System.out.println(count);
View Full Code Here
TOP

Related Classes of org.fnlp.nlp.cn.tag.CWSTagger

  • org.apache.commons.cli.BasicParser
  • org.apache.commons.cli.CommandLine
  • org.apache.commons.cli.HelpFormatter
  • org.apache.commons.cli.Options
  • org.fnlp.app.keyword.WordExtract
  • org.fnlp.demo.nlp.ChineseWordSegmentation
  • org.fnlp.demo.nlp.KeyWordExtraction
  • org.fnlp.demo.nlp.PartsOfSpeechTag
  • org.fnlp.demo.nlp.tc.TextClassificationBasedOnBayes
  • org.fnlp.demo.nlp.tc.TextClassificationBasedOnKNN
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.