Package org.ictclas4j.bean

Examples of org.ictclas4j.bean.SegNode

587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
            else
              break;
          }
        }
        if (end > start) {
          SegNode newsn = new SegNode();
          newsn.setRow(sns.get(start).getRow());
          newsn.setCol(sns.get(end - 1).getCol());
          newsn.setPos(pos);
          newsn.setWord(unknownFlags);
          newsn.setSrcWord(srcWord);
          double value = computePossibility(start, end - start + 1, sns);
          newsn.setValue(value);
          segGraph.insert(newsn, true);
        }
      }
    }
  }
View Full Code Here
603
604
605
606
607
608
609
610
611
612
613
    }
  }

  private int getBestTag(ArrayList<SegNode> sns, int index) {
    if (sns != null && index >= 0 && index < sns.size()) {
      SegNode sn = sns.get(index);
      return getBestTag(sn);

    }

    return -1;
View Full Code Here
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
  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
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
    String temp=null;
    char[] pos = new char[2];
    if (wrList != null && wrList.size() > 0) {
      result = "";
      for (int i = 0; i < wrList.size(); i++) {
        SegNode sn = wrList.get(i);
        if (sn.getPos() != POSTag.SEN_BEGIN && sn.getPos() != POSTag.SEN_END) {
          int tag = Math.abs(sn.getPos());
          pos[0] = (char) (tag / 256);
          pos[1] = (char) (tag % 256);
          temp=""+pos[0];
          if(pos[1]>0)
            temp+=""+pos[1];
          result += sn.getSrcWord() + "/" + temp + " ";
        }
      }
    }

    return result;
View Full Code Here
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
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
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
    if (sns != null && coreDict != null && unknownDict != null && context != null) {
      int i = 0;
      String curWord = null;

      for (; i < sns.size(); i++) {
        SegNode sn = sns.get(i);
        sn.setAllPos(null);
        curWord = sn.getSrcWord();
        // if (tagType == Utility.TAG_TYPE.TT_NORMAL ||
        // !unknownDict.isExist(sn.getWord(), 44)) {
        //
        // }

        if (tagType != Utility.TAG_TYPE.TT_NORMAL) {

          // ��ȫ���ַ����ɰ�ǵ��ַ�
          if (tagType == Utility.TAG_TYPE.TT_TRANS_PERSON && i > 0) {
            String prevWord = sns.get(i - 1).getSrcWord();
            if (Utility.charType(prevWord) == Utility.CT_CHINESE) {
              if (".".equals(curWord))
                curWord = "��";
              else if ("-".equals(curWord))
                curWord = "��";
            }
          }

          // ��unknownDict�ʵ���л�ȡ��ǰ�����д���
          ArrayList<WordItem> wis = unknownDict.getHandle(curWord);
          for (int j = 0; wis != null && j < wis.size(); j++) {
            WordItem wi = wis.get(j);
            int tag = wi.getHandle();
            double freq = -Math.log((1 + wi.getFreq()));
            freq += Math.log((context.getFreq(0, wi.getHandle()) + wis.size() + 1));
            POS pos = new POS(tag, freq);
            sn.addPos(pos);
          }

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

          else if (Utility.SENTENCE_END.equals(curWord))
            sn.addPos(new POS(101, 0));
          else {
            int nFreq = 0;
            wis = coreDict.getHandle(curWord);
            if (wis != null) {
              for (WordItem wi : wis)
                nFreq += wi.getFreq();

              if (wis.size() > 0) {
                double freq = -Math.log((double) (1 + nFreq));
                freq += Math.log((double) (context.getFreq(0, 0) + wis.size()));
                sn.addPos(new POS(0, freq));
              }
            }
          }
        } else {
          if (sn.getPos() > 0) {
            int tag = sn.getPos();
            double value = -Math.log(sn.getValue());
            value += Math.log(context.getFreq(0, tag));
            if (value < 0)
              value = 0;
            sn.addPos(new POS(tag, value));
          } else {
            if (sn.getPos() < 0) {
              sn.setPos(-sn.getPos());
              sn.addPos(new POS(-sn.getPos(), sn.getValue()));
            }
            ArrayList<WordItem> wis = coreDict.getHandle(curWord);
            if (wis != null) {
              for (WordItem wi : wis) {
                int tag = wi.getHandle();
                double value = -Math.log(1 + wi.getFreq());
                value += Math.log(context.getFreq(0, tag) + wis.size());
                sn.addPos(new POS(tag, value));
              }
            }
          }
        }

        if (sn.getAllPos() == null)
          guessPos(tagType, sn);
       
        // ���һ���ʽڵ��Ӧ��allPosΪnull����˵�����޷������ɴ�
        // ���Ĵ�������һ���ʵĴ���
        if (i - 1 >= 0 && sns.get(i - 1).getPosSize() == -1) {
          if (sn.getPosSize() > 0) {
            POS pos = new POS(sn.getAllPos().get(0).getTag(), 0);
            sns.get(i - 1).addPos(pos);
          }
        }
      }

      // ���һ��������
      SegNode last = sns.get(i - 1);
      if (last != null) {
        SegNode sn = new SegNode();
        int tag = 0;
        if (tagType != Utility.TAG_TYPE.TT_NORMAL)
          tag = 101;
        else
          tag = 1;
        POS pos = new POS(tag, 0);
        sn.addPos(pos);
        sns.add(sn);
      }
    }
  }
View Full Code Here
134
135
136
137
138
139
140
141
142
143
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
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
    if (sns != null && coreDict != null && unknownDict != null && context != null) {
      int i = 0;
      String curWord = null;

      for (; i < sns.size(); i++) {
        SegNode sn = sns.get(i);
        sn.setAllPos(null);
        curWord = sn.getSrcWord();
        int gbkID = sn.getGbkID();// dictLib.getGBKID(curWord);
        // if (tagType == Utility.TAG_TYPE.TT_NORMAL ||
        // !unknownDict.isExist(sn.getWord(), 44)) {
        //
        // }

        if (tagType != Utility.TAG_TYPE.TT_NORMAL) {

          // ��ȫ���ַ����ɰ�ǵ��ַ�
          if (tagType == Utility.TAG_TYPE.TT_TRANS_PERSON && i > 0) {
            String prevWord = sns.get(i - 1).getSrcWord();
            if (Utility.charType(prevWord) == Utility.CT_CHINESE) {
              if (".".equals(curWord))
                curWord = "��";
              else if ("-".equals(curWord))
                curWord = "��";
            }
          }

          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));
              }
            }
          }
        }

        if (sn.getAllPos() == null)
          guessPos(tagType, sn);

        // ���һ���ʽڵ��Ӧ��allPosΪnull����˵�����޷������ɴ�
        // ���Ĵ�������һ���ʵĴ���,���ǽ�����ʶ��ĩ##ĩ������
        if (i - 1 >= 0 && sns.get(i - 1).getPosSize() == -1) {
          if (sn.getPosSize() > 0) {
            Pos pos = sn.getAllPos().get(0).getPos();
            int ipos = pos.getTag() == POSTag.SEN_END ? POSTag.UNKNOWN : pos.getTag();
            AdjoiningPos apos = new AdjoiningPos( ipos , 0);
            sns.get(i - 1).addPos(apos);
          }
        }
      }

      // ���һ��������
      SegNode last = sns.get(i - 1);
      if (last != null) {
        SegNode sn = new SegNode();
        int tag = 0;
        if (tagType != Utility.TAG_TYPE.TT_NORMAL)
          tag = 101;
        else
          tag = 1;
        AdjoiningPos pos = new AdjoiningPos( tag, 0);
        sn.addPos(pos);
        sns.add(sn);
      }
    }
  }
View Full Code Here
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
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
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
    if (sns != null && coreDict != null && unknownDict != null && context != null) {
      int i = 0;
      String curWord = null;

      for (; i < sns.size(); i++) {
        SegNode sn = sns.get(i);
        sn.setAllPos(null);
        curWord = sn.getSrcWord();
        // if (tagType == Utility.TAG_TYPE.TT_NORMAL ||
        // !unknownDict.isExist(sn.getWord(), 44)) {
        //
        // }

        if (tagType != Utility.TAG_TYPE.TT_NORMAL) {

          // ��ȫ���ַ����ɰ�ǵ��ַ�
          if (tagType == Utility.TAG_TYPE.TT_TRANS_PERSON && i > 0) {
            String prevWord = sns.get(i - 1).getSrcWord();
            if (Utility.charType(prevWord) == Utility.CT_CHINESE) {
              if (".".equals(curWord))
                curWord = "��";
              else if ("-".equals(curWord))
                curWord = "��";
            }
          }

          // ��unknownDict�ʵ���л�ȡ��ǰ�����д���
          ArrayList<WordItem> wis = unknownDict.getHandle(curWord);
          for (int j = 0; wis != null && j < wis.size(); j++) {
            WordItem wi = wis.get(j);
            int tag = wi.getHandle();
            double freq = -Math.log((1 + wi.getFreq()));
            freq += Math.log((context.getFreq(0, wi.getHandle()) + wis.size() + 1));
            POS pos = new POS(tag, freq);
            sn.addPos(pos);
          }

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

          else if (Utility.SENTENCE_END.equals(curWord))
            sn.addPos(new POS(101, 0));
          else {
            int nFreq = 0;
            wis = coreDict.getHandle(curWord);
            if (wis != null) {
              for (WordItem wi : wis)
                nFreq += wi.getFreq();

              if (wis.size() > 0) {
                double freq = -Math.log((double) (1 + nFreq));
                freq += Math.log((double) (context.getFreq(0, 0) + wis.size()));
                sn.addPos(new POS(0, freq));
              }
            }
          }
        } else {
          if (sn.getPos() > 0) {
            int tag = sn.getPos();
            double value = -Math.log(sn.getValue());
            value += Math.log(context.getFreq(0, tag));
            if (value < 0)
              value = 0;
            sn.addPos(new POS(tag, value));
          } else {
            if (sn.getPos() < 0) {
              sn.setPos(-sn.getPos());
              sn.addPos(new POS(-sn.getPos(), sn.getValue()));
            }
            ArrayList<WordItem> wis = coreDict.getHandle(curWord);
            if (wis != null) {
              for (WordItem wi : wis) {
                int tag = wi.getHandle();
                double value = -Math.log(1 + wi.getFreq());
                value += Math.log(context.getFreq(0, tag) + wis.size());
                sn.addPos(new POS(tag, value));
              }
            }
          }
        }

        if (sn.getAllPos() == null)
          guessPos(tagType, sn);
       
        // ���һ���ʽڵ��Ӧ��allPosΪnull����˵�����޷������ɴ�
        // ���Ĵ�������һ���ʵĴ���
        if (i - 1 >= 0 && sns.get(i - 1).getPosSize() == -1) {
          if (sn.getPosSize() > 0) {
            POS pos = new POS(sn.getAllPos().get(0).getTag(), 0);
            sns.get(i - 1).addPos(pos);
          }
        }
      }

      // ���һ��������
      SegNode last = sns.get(i - 1);
      if (last != null) {
        SegNode sn = new SegNode();
        int tag = 0;
        if (tagType != Utility.TAG_TYPE.TT_NORMAL)
          tag = 101;
        else
          tag = 1;
        POS pos = new POS(tag, 0);
        sn.addPos(pos);
        sns.add(sn);
      }
    }
  }
View Full Code Here
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
            nPos = j;
            personName = "";
            // Get the possible person name
            while (nPos < j + patterns[k].length()) {
              SegNode sn = sns.get(nPos);
              if (sn.getPos() < 4
                  && unknownDict.getFreq(sn.getWord(), sn.getPos()) < Utility.LITTLE_FREQUENCY)
                personName += sn.getWord();
              nPos += 1;
            }
            if ("CDCD".equals(patterns[k])) {
              if (GetForeignCharCount(personName) > 0)
                j += patterns[k].length() - 1;
              continue;
            }

            SegNode usn = new SegNode();
            usn.setRow(sns.get(j).getRow());
            usn.setCol(sns.get(j + patterns[k].length() - 1).getCol());
            usn.setWord(unknownFlags);
            usn.setSrcWord(personName);
            double value = -Math.log(factor[k]) + computePossibility(j, patterns[k].length(), sns);
            usn.setPos(pos);
            usn.setValue(value);
            segGraph.insert(usn, true);

            j += patterns[k].length();
            bMatched = true;
          }
View Full Code Here
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
            else
              break;
          }
        }
        if (end > start) {
          SegNode newsn = new SegNode();
          newsn.setRow(sns.get(start).getRow());
          newsn.setCol(sns.get(end - 1).getCol());
          newsn.setPos(pos);
          newsn.setWord(unknownFlags);
          newsn.setSrcWord(srcWord);
          double value = computePossibility(start, end - start + 1, sns);
          newsn.setValue(value);
          segGraph.insert(newsn, true);
        }
      }
    }
  }
View Full Code Here
604
605
606
607
608
609
610
611
612
613
614
    }
  }

  private int getBestTag(ArrayList<SegNode> sns, int index) {
    if (sns != null && index >= 0 && index < sns.size()) {
      SegNode sn = sns.get(index);
      return getBestTag(sn);

    }

    return -1;
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.