Package org.cleartk.classifier

Examples of org.cleartk.classifier.Feature

82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
      for (BaseToken token : tokens) {
        ++outcomeIndex;

        // extract token features
        List<Feature> features = new ArrayList<Feature>();
        features.add(new Feature(token.getCoveredText()));
        features.add(new Feature("PartOfSpeech", token.getPartOfSpeech()));

        // extract previous classification features
        for (int i = this.nPreviousClassifications; i > 0; --i) {
          int index = outcomeIndex - i;
          String previousOutcome = index < 0 ? "O" : outcomes.get(index);
          features.add(new Feature("PreviousOutcome_" + i, previousOutcome));
        }

        // extract length of Modifier that is currently being created (if any)
        // int length = 0;
        // for (int i = outcomeIndex - 1; i > 0 && !"O".equals(outcomes.get(i)); --i) {
View Full Code Here
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
125
126
127
128
129
130
131
       
    // Extract features between
    List<TerminalTreebankNode> headList = this.extractPhraseHeadByTreenode(jCas, JCasUtil.selectCovered(jCas, TreebankNode.class, arg1.getEnd(), arg2.getBegin()));

    if(headList.size() > 0) {
      features.add(new Feature("PhraseChunk_Between_FirstHead", headList.get(0).getNodeValue()));
      features.add(new Feature("PhraseChunk_Between_LastHead", headList.get(headList.size()-1).getNodeValue()));
     
      if(headList.size() >= 2) {
        String inBetweenValue = "";
        for(int i=1;i<headList.size()-1;i++) {
          if(i>1)
            inBetweenValue += "_";
          inBetweenValue += headList.get(i).getNodeValue();
        }
        features.add(new Feature("PhraseChunk_Between_BetweenHeads", inBetweenValue));
      }
    }
   
    // Extract feature before M1
    headList = this.extractPhraseHeadByTreenode(jCas, JCasUtil.selectPreceding(jCas, TreebankNode.class, arg1, 20));

    boolean isFirst = false;
    for(int i=headList.size()-1;i>=0;i--) {
      if(headList.get(i).getEnd() < arg1.getBegin()) {
        if(!isFirst) {
        features.add(new Feature("PhraseChunk_Before_FirstHead", headList.get(i).getNodeValue()));
        isFirst = true;
        }
        else {
          features.add(new Feature("PhraseChunk_Before_SecondHead", headList.get(i).getNodeValue()));
          break;
        }
      }
    }
   
    // Extract feature after M2
    headList = this.extractPhraseHeadByTreenode(jCas, JCasUtil.selectFollowing(jCas, TreebankNode.class, arg2, 20));
   
   
    isFirst = false;
    for(int i=0;i<headList.size();i++) {
      if(headList.get(i).getBegin() > arg2.getEnd() ) {
        if(!isFirst) {
          features.add(new Feature("PhraseChunk_After_FirstHead", headList.get(i).getNodeValue()));
          isFirst = true;
        }
        else {
          features.add(new Feature("PhraseChunk_After_SecondHead", headList.get(i).getNodeValue()));
          break;
        }
      }
    }
   
View Full Code Here
258
259
260
261
262
263
264
265
266
267
268
269
270
271
        // features from previous classifications
        int nPreviousClassifications = 2;
        for (int i = nPreviousClassifications; i > 0; --i) {
          int index = tokenIndex - i;
          String previousOutcome = index < 0 ? "O" : outcomes.get(index);
          features.add(new Feature("PreviousOutcome_" + i, previousOutcome));
        }
        //add segment ID as a features:
        features.add(new Feature("SegmentID", segment.getId()));

        // features from dominating parse tree
        //        for(SimpleFeatureExtractor extractor : this.parseFeatureExtractors){
        BaseToken startToken = token;
        for(int i = tokenIndex-1; i >= 0; --i){
View Full Code Here
59
60
61
62
63
64
65
66
67
68
69
70
71
   
    List<EventMention> events = JCasUtil.selectCovered(jCas, EventMention.class, begin, end);
    List<TimeMention> times   = JCasUtil.selectCovered(jCas, TimeMention.class, begin, end);
    eventsInBetween = events==null? 0: events.size();
    timesInBetween  = times==null? 0: times.size();
    feats.add(new Feature("NumOfEvents_InBetween", eventsInBetween));
    feats.add(new Feature("NumOfTimes_InBetween", timesInBetween));
    feats.add(new Feature("NumOfEventsAndTimes_InBetween", timesInBetween+eventsInBetween));
   
//    //print long distances
//    if (eventsInBetween >= 200){
//      System.out.println("++++++++++Long Distance Relation in "+ ViewURIUtil.getURI(jCas).toString() + "+++++++");
//      System.out.println("["+arg1.getCoveredText()+"] "+ jCas.getDocumentText().substring(arg1.getEnd(), arg2.getBegin()) +" ["+arg2.getCoveredText()+"]");
View Full Code Here
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
      {
    List<Feature> feats = new ArrayList<Feature>();
   
    TreebankNode domNode = AnnotationTreeUtils.annotationNode(jcas, begin, end);
    if(domNode != null){
      feats.add(new Feature("DominatingTreeCat", domNode.getNodeType()));
      if(domNode.getNodeTags() != null){
        for(int ind = 0; ind < domNode.getNodeTags().size(); ind++){
          String tag = domNode.getNodeTags(ind);
          if(tag.equals("TMP")){
            feats.add(new Feature("DominatingTmpTag", tag));
          }
        }
      }
      TreebankNode parent = domNode.getParent();
      if(parent != null){
        feats.add(new Feature("DominatingTreeParent", parent.getNodeType()));
        do{
          if(parent.getNodeTags() != null){
            for(int ind = 0; ind < parent.getNodeTags().size(); ind++){
              String tag = parent.getNodeTags(ind);
//              if(tag.equals("TMP")){
                feats.add(new Feature("DominatingAncestorTmpTag", tag));
//              }
            }
          }
          parent = parent.getParent();
        }while(parent != null);
      }
     
      if(domNode.getLeaf()){
        feats.add(new Feature("DominatingIsLeaf"));
      }else{
        StringBuffer buffer = new StringBuffer();
        for(int i = 0; i < domNode.getChildren().size(); i++){
          buffer.append(domNode.getChildren(i).getNodeType());
          buffer.append("_");
          feats.add(new Feature("DominatingChildBag" + domNode.getChildren(i).getNodeType()));
        }
        feats.add(new Feature("DominatingProduction", buffer.toString()));
      }
      if(domNode.getBegin() == begin && domNode.getEnd() == end){
        feats.add(new Feature("DominatingExactMatch"));
      }
    }
    return feats;
  }
View Full Code Here
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
  }

  public List<Feature> extract(BaseToken token) {
    List<Feature> features = Lists.newArrayList();
    Collection<Predicate> predicates = this.tokenPredicateMap.get(token);
    features.add(new Feature("Predicate", !predicates.isEmpty()));
    for (Predicate predicate : predicates) {
      features.add(new Feature("Predicate_Lex", predicate.getCoveredText()));
    }
    for (SemanticArgument argument : this.tokenArgumentMap.get(token)) {
      SemanticRoleRelation relation = argument.getRelation();
      String category = relation.getCategory();
      features.add(new Feature("Argument", category));
      String predicateText = relation.getPredicate().getCoveredText();
      features.add(new Feature("Argument_Lex", String.format("%s_%s", category, predicateText)));
    }
    return features;
  }
View Full Code Here
44
45
46
47
48
49
50
51
52
53
    {
      return features;
    }

    LinkedList<ConllDependencyNode> node1ToNode2Path = DependencyParseUtils.getPathBetweenNodes(node1, node2);
    features.add(new Feature("dependency_path", DependencyParseUtils.pathToString(node1ToNode2Path)));
    features.add(new Feature("dependency_path_length", node1ToNode2Path.size()));//add path length as a feature

    return features;
  }
View Full Code Here
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
   
    //2 get DateX
    if (sentList != null && !sentList.isEmpty()){
      for(Sentence sent : sentList) {
        for (@SuppressWarnings("unused") DateAnnotation date : JCasUtil.selectCovered(view, DateAnnotation.class, sent)) {
          Feature indicator = new Feature("DateXNearby", this.name);
          features.add(indicator);
          break;
        }
      }
    }
   
  //3 get Measurement
    if (sentList != null && !sentList.isEmpty()){
      for(Sentence sent : sentList) {
        for (@SuppressWarnings("unused") MeasurementAnnotation date : JCasUtil.selectCovered(view, MeasurementAnnotation.class, sent)) {
          Feature indicator = new Feature("MeasurementNearby", "measure");
          features.add(indicator);
          break;
        }
      }
    }
   
  //4 get number
    if (sentList != null && !sentList.isEmpty()){
      for(Sentence sent : sentList) {
        for (@SuppressWarnings("unused") NumToken date : JCasUtil.selectCovered(view, NumToken.class, sent)) {
          Feature indicator = new Feature("NumTokenNearby", "NumToken");
          features.add(indicator);
          break;
        }
      }
    }
View Full Code Here
70
71
72
73
74
75
76
77
78
79
80
81
82
83
              verbDistMap.put(Math.abs(wt.getBegin() - annotation.getBegin()), wt);
            }
          }
        }
        for (Map.Entry<Integer, WordToken> entry : verbDistMap.entrySet()) {
          Feature feature = new Feature(this.name+"_token", entry.getValue().getCoveredText());
          features.add(feature);
          //logger.info("found nearby closest verb: "+ entry.getValue().getCoveredText() + " POS:" + entry.getValue().getPartOfSpeech());
          Feature posfeature = new Feature(this.name, entry.getValue().getPartOfSpeech());
          features.add(posfeature);
          break;             
        }
      }
     
View Full Code Here
57
58
59
60
61
62
63
64
65
66
67
     
      // print out totals:
      for(String typeId : typeCounts.keySet()){
        String featName = "eventTypeID_"+typeId;
        eventTypes.add(featName);
        features.add(new Feature(featName, typeCounts.get(typeId)));       
      }
     
      // TO print out just the types without counts:
//      for(String typeId : typeCounts.keySet()){
//        features.add(new Feature("arg1EntityTypeID_", typeId));
View Full Code Here
TOP

Related Classes of org.cleartk.classifier.Feature

  • org.apache.ctakes.assertion.attributes.features.GenericFeaturesExtractor
  • org.apache.ctakes.assertion.attributes.features.HistoryFeaturesExtractor
  • org.apache.ctakes.assertion.attributes.features.SubjectFeaturesExtractor
  • org.apache.ctakes.assertion.medfacts.cleartk.AssertionCleartkAnalysisEngine
  • org.apache.ctakes.assertion.medfacts.cleartk.extractors.AboveLeftFragmentExtractor
  • org.apache.ctakes.assertion.medfacts.cleartk.extractors.AboveRightFragmentExtractor
  • org.apache.ctakes.assertion.medfacts.cleartk.extractors.BrownClusterFeatureFunction
  • org.apache.ctakes.assertion.medfacts.cleartk.extractors.ContextWordWindowExtractor
  • org.apache.ctakes.assertion.medfacts.cleartk.extractors.DependencyWordsFragmentExtractor
  • org.apache.ctakes.assertion.medfacts.cleartk.extractors.FedaFeatureFunction
    • 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.