Package org.ictclas4j.bean

Examples of org.ictclas4j.bean.SegNode

723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
  private double computePossibility(int startPos, int length, ArrayList<SegNode> sns) {
    double retValue = 0, posPoss;

    if (sns != null && unknownDict != null && context != null) {
      for (int i = startPos; i < startPos + length && sns != null; i++) {
        SegNode sn = sns.get(i);
        int bestTag = getBestTag(sn);
        if (bestTag != -1) {
          int freq = unknownDict.getFreq(sn.getSrcWord(), bestTag);
          posPoss = Math.log((double) (context.getFreq(0, sn.getPos()) + 1));
          posPoss += -Math.log((double) (freq + 1));
          retValue += posPoss;
        }
      }
    }
View Full Code Here
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
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
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
   * @param atoms
   * @return
   */
  public static SegGraph generate(ArrayList<Atom> atoms,Dictionary dict) {
    SegGraph segGraph = null;
    SegNode sn = null;
    Atom atom = null;

    if (atoms != null && atoms.size() > 0 && dict != null) {
      segGraph = new SegGraph();
      for (int i = 0; i < atoms.size(); i++) {
        atom = atoms.get(i);
        String word = atom.getWord();
        if (atom.getPos() == Utility.CT_CHINESE)
          sn = new SegNode(i, i + 1, 0,0, atom.getWord());
        else {
          double value = Utility.MAX_FREQUENCE;
          int pos = 0;

          switch (atom.getPos()) {
          case Utility.CT_INDEX:
          case Utility.CT_NUM:
            pos = -POSTag.NUM;// 'm'*256
            word = Utility.UNKNOWN_NUM;
            value = 0;
            break;
          case Utility.CT_DELIMITER:
            pos = POSTag.PUNC;// 'w'*256;
            break;
          case Utility.CT_LETTER:
            pos = -POSTag.NOUN_LETTER;//
            value = 0;
            word = Utility.UNKNOWN_LETTER;
            break;
          case Utility.CT_SINGLE:// 12021-2129-3121
            if (Utility.getCharCount("+-1234567890", atom.getWord()) == atom.getLen()) {
              pos = -POSTag.NUM;// 'm'*256
              word = Utility.UNKNOWN_NUM;
            } else {
              pos = -POSTag.NOUN_LETTER;//
              word = Utility.UNKNOWN_LETTER;
            }
            value = 0;
            break;
          default:
            pos = atom.getPos();// '?'*256;
            break;
          }

          sn = new SegNode(i, i + 1,pos, value , word);
        }

        sn.setSrcWord(atom.getWord());
        segGraph.insert(sn, true);
      }

      String word = null;
      for (int i = 0; i < atoms.size(); i++) {
        int j = i + 1;
        word = atoms.get(i).getWord();
        // ����ǡ��·ݡ�����Ҫ�ָ�
        boolean flag = false;
        if (j < atoms.size()) {
          Atom a2 = atoms.get(j);
          if ("��".equals(word) && "��".equals(a2.getWord())) {
            segGraph.delete(i, j);
            segGraph.delete(i + 1, j + 1);
            word += a2.getWord();
            flag = true;
            j++;
          }
        }

        WordItem wi = null;
        for (; j <= atoms.size(); j++) {
          int totalFreq = 0;
          wi = dict.getMaxMatch(word);
          if (wi != null) {
            // find it
            if (word.equals(wi.getWord())) {
              ArrayList<WordItem> wis = dict.getHandle(word);
              for (WordItem w : wis)
                totalFreq += w.getFreq();

              // 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(word.substring(1)) != -1)
                    break;
                }
              }
              // ֻ��һ���Դʣ�������
              SegNode sg = null;
              if (wis.size() == 1)
                sg = new SegNode(i, j,wis.get(0).getHandle(),totalFreq , word);
              else
                sg = new SegNode(i, j, 0,totalFreq , word);
           
              segGraph.insert(sg, true);

            }
            if (flag)
View Full Code Here
150
151
152
153
154
155
156
157
158
159
160
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
    if (seg != null && dict != null && biDict != null) {
      segGraph = new SegGraph();
      ArrayList<SegNode> sgs = seg.getSnList();

      for (int i = 0; sgs != null && i < sgs.size(); i++) {
        SegNode sg = sgs.get(i);
        if (sg.getPos() >= 0)
          curFreq = sg.getValue();
        else
          curFreq = dict.getFreq(sg.getWord(), 2);

        // �õ�������ֵ�͸���ֵ��ȵ�����Ԫ��
        ArrayList<SegNode> nextSgs = seg.getNextElements(i);
        for (SegNode graph : nextSgs) {
          String twoWords = sg.getWord();
          twoWords += Utility.WORD_SEGMENTER;
          twoWords += graph.getWord();

          // ��������������֮���ƽ��ֵ
          // -log{a*P(Ci-1)+(1-a)P(Ci|Ci-1)} Note 0<a<1
          int twoFreq = biDict.getFreq(twoWords, 3);
          double temp = (double) 1 / Utility.MAX_FREQUENCE;
          double value = smoothParam * (1 + curFreq) / (Utility.MAX_FREQUENCE + 80000);
          value += (1 - smoothParam) * ((1 - temp) * twoFreq / (1 + curFreq) + temp);
          value = -Math.log(value);

          if (value < 0)
            value += sg.getValue();

          SegNode sg2 = new SegNode();
          // �ָ���@ǰ�Ĵ��������е�λ��
          int wordIndex = getWordIndex(sgs, sg);
          sg2.setRow(wordIndex);

          // �ָ���@��Ĵ��������е�λ��
          wordIndex = getWordIndex(sgs, graph);
          sg2.setCol(wordIndex);
          sg2.setWord(twoWords);
          sg2.setPos(sg.getPos());
          sg2.setValue(value);
          segGraph.insert(sg2, false);
        }
      }
    }
    return segGraph;
View Full Code Here
27
28
29
30
31
32
33
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
62
63
64
65
  public SegGraph(ArrayList<SegNode> snList) {
    this.snList = snList;
  }

  public SegNode getElement(int row, int col) {
    SegNode result = new SegNode();
    result.setValue(  Utility.INFINITE_VALUE );
    // if (row > m_nRow || col > m_nCol)
    // return null;

    int index = 0;
    if (snList != null) {
      if (isRowFirst) {
        for (int i = 0; i < snList.size(); i++, index++) {
          SegNode sg = snList.get(i);
          if (row != -1 && sg.getRow() < row || col != -1 && sg.getRow() == row && sg.getCol() < col)
            continue;
          else
            break;
        }
      } else {
        for (int i = 0; i < snList.size(); i++, index++) {
          SegNode sg = snList.get(i);
          if (col != -1 && sg.getCol() < col || row != -1 && sg.getCol() == col && sg.getRow() < row)
            continue;
          else
            break;
        }
      }

      // Find it and return the value
      if (index < snList.size()) {
        SegNode sg = snList.get(index);
        if ((sg.getRow() == row || row == -1) && (sg.getCol() == col || col == -1))
          result = sg;
      }
    }
    return result;
View Full Code Here
75
76
77
78
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
105
106
    if (sg != null) {
      if (snList == null)
        snList = new ArrayList<SegNode>();

      int i = 0;
      SegNode sgTemp = null;
      if (isRowFirst) {
        for (i = 0; i < snList.size(); i++) {
          sgTemp = snList.get(i);
          if (sgTemp.getRow() < sg.getRow() || sgTemp.getRow() == sg.getRow()
              && sgTemp.getCol() < sg.getCol())
            continue;
          else
            break;
        }
      } else {
        for (i = 0; i < snList.size(); i++) {
          sgTemp = snList.get(i);
          if (sgTemp.getCol() < sg.getCol() || sgTemp.getCol() == sg.getCol()
              && sgTemp.getRow() < sg.getRow())
            continue;
          else
            break;
        }
      }

      if (sgTemp != null && sgTemp.getRow() == sg.getRow() && sgTemp.getCol() == sg.getCol())
        sgTemp = sg;
      else if (i > 0)
        snList.add(i - 1, sg);

    }
View Full Code Here
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
    ArrayList<SegNode> result = null;

    if (snList != null && snList.size() > 0 && curIndex >= 0) {
      result = new ArrayList<SegNode>();
      for (int i = 0; i < snList.size(); i++) {
        SegNode sg = snList.get(i);
        if (isColFirst) {
          if (sg.getCol() == curIndex)
            result.add(sg);
        } else {
          if (sg.getRow() == curIndex)
            result.add(sg);
        }

      }
    }
View Full Code Here
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
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
   * @param isRowFirst
   *            �Ƿ�������ԭ��
   * @return �������ɹ�����True,���򷵻�False
   */
  public boolean insert(SegNode graph, boolean isRowFirst) {
    SegNode sg = null;
    if (snList == null)
      snList = new ArrayList<SegNode>();

    if (graph != null) {
      if (snList.size() == 0) {
        snList.add(graph);
        return true;
      }

      for (int i = 0; i < snList.size(); i++) {
        sg = snList.get(i);

        if (isRowFirst) {
          // �����һ���ڵ�
          if (i == snList.size() - 1) {
            if (graph.getRow() > sg.getRow()
                || (graph.getRow() == sg.getRow() && graph.getCol() > sg.getCol()))
              snList.add(graph);
            else {
              if (graph.getCol() == sg.getCol()) {
                snList.set(i, graph);
              } else
                snList.add(i, graph);
            }
            return true;
          }

          if (graph.getRow() > sg.getRow() || (graph.getRow() == sg.getRow() && graph.getCol() > sg.getCol()))
            continue;
          else {
            if (graph.getRow()==sg.getRow() && graph.getCol() == sg.getCol()) {
              snList.set(i, graph);
            } else
              snList.add(i, graph);

            return true;
          }
        } else {
          // �����һ���ڵ�
          if (i == snList.size() - 1) {
            if (graph.getCol() > sg.getCol()
                || (graph.getCol() == sg.getCol() && graph.getRow() > sg.getRow()))
              snList.add(graph);
            else {
              if (graph.getRow()==sg.getRow() && graph.getCol() == sg.getCol()) {
                snList.set(i, graph);
              } else
                snList.add(i, graph);
            }
            return true;
          }

          if (graph.getCol() > sg.getCol() || (graph.getCol() == sg.getCol() && graph.getRow() > sg.getRow()))
            continue;
          else {
            if (graph.getRow() == sg.getRow()) {
              snList.set(i, graph);
            } else
              snList.add(i, graph);

            return true;
View Full Code Here
215
216
217
218
219
220
221
222
223
224
225
    return false;
  }

  public SegNode delete(int row, int col) {
    SegNode result = null;

    if (snList != null && snList.size() > 0) {
      for (SegNode sn : snList) {
        if (sn.getRow() == row && sn.getCol() == col) {
          snList.remove(sn);
View Full Code Here
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
  public ArrayList<SegNode> getNextElements(int curIndex) {
    ArrayList<SegNode> result = null;

    if (snList != null && snList.size() > 0 && curIndex >= 0 && curIndex < snList.size()) {
      result = new ArrayList<SegNode>();
      SegNode curSg = snList.get(curIndex);

      for (int i = curIndex + 1; i < snList.size(); i++) {
        SegNode sg = snList.get(i);
        if (sg.getRow() == curSg.getCol())
          result.add(sg);

      }
    }
    return result;
View Full Code Here
282
283
284
285
286
287
288
289
290
291
292
  public int getMaxRow() {
    int result = -1;

    if (snList != null && snList.size() > 0) {
      int size = snList.size();
      SegNode sn = snList.get(size - 1);
      result = sn.getRow();
    }

    return result;
  }
View Full Code Here
TOP

Related Classes of org.ictclas4j.bean.SegNode

  • org.ictclas4j.segment.AdjustSeg
  • org.ictclas4j.segment.GraphGenerate
  • org.ictclas4j.segment.PosTagger
  • org.ictclas4j.segment.SegGraph
  • org.ictclas4j.segment.Segment
    • 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.