Package org.ictclas4j.bean

Examples of org.ictclas4j.bean.SegAtom

87
88
89
90
91
92
93
94
95
96
97
              midResult = outputResult(adjResult);
            }
            break;
          }
        } else {
          SegAtom atom = new SegAtom(sen.getContent());
          SegAtom[] atoms = new SegAtom[1];
          atoms[0] = atom;
          midResult = new SegResult();
          midResult.setRawContent(sen.getContent());
          midResult.setAtoms(atoms);
View Full Code Here
188
189
190
191
192
193
194
195
196
197
198
      ArrayList<SegAtom> saList = new ArrayList<SegAtom>();
      for (int i = 0; i < wrList.size(); i++) {

        SegNode sn = wrList.get(i);
        if (sn.getPos() != POSTag.SEN_BEGIN && sn.getPos() != POSTag.SEN_END) {
          SegAtom sa =sn.toSegAtom();
          saList.add(sa);
        }
      }

      SegAtom[] atoms = new SegAtom[saList.size() - 1];
View Full Code Here
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
            flag = true;
            j++;
          }
        }

        SegAtom sa = null;
        String word = words.toString();
        int gbkID = dictLib.getGBKID(word);
        int wordMaxLen = dict.getWordMaxLen(word, gbkID);
        for (; j <= atoms.size() && word.length() < wordMaxLen; j++) {
          word = words.toString();
          sa = dict.getSegAtom(word, gbkID);
          if (sa != null) {
            // 1���ڣ�1999��ĩ
            // if (word.length() == 2 && segGraph.getSize() > 0) {
            // SegNode g2 = segGraph.getLast();
            // if (Utility.isAllNum(g2.getWord()) ||
            // Utility.isAllChinese(g2.getWord())
            // && (g2.getWord().indexOf("��") == 0 ||
            // g2.getWord().indexOf("��") == 0)) {
            // if ("ĩ���е�ǰ���".indexOf(words.substring(1)) != -1)
            // break;
            // }
            // }
            // ֻ��һ���Դʣ�������
            SegNode sg = null;
            if (sa.getPosCount() == 1) {
              Pos pos = sa.getPos(0);
              sg = new SegNode(i, j, pos.getTag(), sa.getTotalFreq(), word);
            } else
              sg = new SegNode(i, j, 0, sa.getTotalFreq(), word);
            sg.setSrcWord(word);
            sg.setGbkID(gbkID);
            segGraph.insert(sg, true);
          }
View Full Code Here
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
      int wordMaxLen=0;
      if(wis0!=null){
        System.out.println("size:"+i+","+wis0.size()+","+size);
        HashMap<String, SegAtom> wordMap = new HashMap<String, SegAtom>();
        for(int j=0;j<wis0.size();j++){
          SegAtom sa=new SegAtom();
          WordItem0 wi=wis0.get(j);
          sa.setWord(wi.getWord());
          sa.addPos(new Pos(wi.getHandle(),wi.getFreq(),false));
          count++;
          size+=8+wi.getWord().getBytes().length;
         
          //����ͬ�Ĵʣ����кϲ�
          while(j<wis0.size()-1 && wis0.get(j).getWord()!=null && wis0.get(j).getWord().equals(wis0.get(j+1).getWord())){
            wi=wis0.get(j+1);
            sa.addPos(new Pos(wi.getHandle(),wi.getFreq(),false));
            j++;
            size+=8;
          }
         
          wordMap.put(sa.getWord(), sa);
          if(sa.getWord().length()>wordMaxLen)
            wordMaxLen=sa.getWord().length();
         
        }
       
        WordTable wt=new WordTable();
        wt.setWordCount(count);
View Full Code Here
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
          if (sn.getPos() < 0) {
            AdjoiningPos pos = new AdjoiningPos( 0 , 0);
            sn.addPos(pos);
          } else {
            // ��unknownDict�ʵ���л�ȡ��ǰ�����д���
            SegAtom sa = unknownDict.getSegAtom(curWord, gbkID);
            for (int j = 0; sa != null && j < sa.getPosCount(); j++) {
              Pos pos = sa.getPos(j);
              double value = -Math.log((1 + pos.getFreq()));
              value += Math.log((context.getFreq(pos.getTag()) + sa.getPosCount() + 1));
              AdjoiningPos apos = new AdjoiningPos(pos , value);
              sn.addPos(apos);
            }

            if (Utility.SENTENCE_BEGIN.equals(curWord))
              sn.addPos(new AdjoiningPos( 100 , 0));

            else if (Utility.SENTENCE_END.equals(curWord))
              sn.addPos(new AdjoiningPos( 101 , 0));
            else {
              int freq = 0;
              sa = coreDict.getSegAtom(curWord, gbkID);
              if (sa != null) {
                double value = -Math.log((double) (1 + freq));
                value += Math.log((double) (context.getFreq(0) + sa.getPosCount()));
                sn.addPos(new AdjoiningPos( 0 , value));

              }
            }
          }
        } else {
          if (sn.getPos() > 0) {
            int tag = sn.getPos();
            double value = -Math.log(1 + sn.getFreq());
            value += Math.log(1 + context.getFreq(tag));
            if (value < 0)
              value = 0;
            sn.addPos(new AdjoiningPos( tag,  value));
          } else {
            if (sn.getPos() < 0) {
              sn.setPos(-sn.getPos());
              sn.addPos(new AdjoiningPos( -sn.getPos(),  sn.getFreq()));
            }
            SegAtom sa = coreDict.getSegAtom(curWord, gbkID);
            if (sa != null) {
              for (int j = 0; j < sa.getPosCount(); j++) {
                Pos pos = sa.getPos(j);
                double value = -Math.log(1 + pos.getFreq());
                value += Math.log(context.getFreq(pos.getTag()) + sa.getPosCount());
                sn.addPos(new AdjoiningPos(pos , value));
              }
            }
          }
        }
View Full Code Here
TOP

Related Classes of org.ictclas4j.bean.SegAtom

  • org.ictclas4j.segment.GraphGenerate
  • org.ictclas4j.segment.PosTagger
  • org.ictclas4j.segment.Segment
  • org.ictclas4j.test2.DictTest
    • 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.