Package types

Examples of types.Alphabet

19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
public class InternetAds {

  public static void main(String[] args) throws IOException {
    // read in the data.
    ArrayList<ClassificationInstance> allData = (new InternetAdReader(
        new Alphabet(), new Alphabet())).readFile(args[0]);
    StaticUtils.shuffle(allData, 0);
    // randomly split data into training and testing part
    ArrayList<ClassificationInstance>[] tmp = StaticUtils.split(allData,
        200);
    ArrayList<ClassificationInstance> train = tmp[0];
    ArrayList<ClassificationInstance> test = tmp[1];
    Alphabet xA = allData.get(0).xAlphabet;
    Alphabet yA = allData.get(0).yAlphabet;
    System.out.println("num Features = " + allData.get(0).xAlphabet.size());
    LinearClassifier h;
    h = trainAdaBoost(50, train, xA, yA);
    System.out.println("Boost  Train Accuracy = "
        + StaticUtils.computeAccuracy(h, train));
View Full Code Here
131
132
133
134
135
136
137
138
139
140
141
142
    }
  }

  public static void main(String[] args) {
    ArrayList<ClassificationInstance> train = new ArrayList<ClassificationInstance>();
    Alphabet xAlphabet = new Alphabet();
    Alphabet yAlphabet = new Alphabet();
    String[] classes = new String[] { "a", "b" };
    Random r = new Random(10);
    int numFeats = 5;
    double randomFrac = 0.5;
    double missingFrac = 0.5;
View Full Code Here
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
public class PartOfSpeech {

  public static void main(String[] args) throws IOException {
    // read in the data.
    ArrayList<SequenceInstance> allData = (new PartOfSpeechReader(
        new Alphabet(), new Alphabet())).readFile(args[0]);
    StaticUtils.shuffle(allData, 0);
    // randomly split data into training and testing part
    ArrayList<SequenceInstance>[] tmp = StaticUtils.splitS(allData, 150);
    ArrayList<SequenceInstance> train = tmp[0];
    ArrayList<SequenceInstance> test = tmp[1];
    Alphabet xA = allData.get(0).xAlphabet;
    Alphabet yA = allData.get(0).yAlphabet;
    System.out.println("num Features = " + allData.get(0).xAlphabet.size());
    LinearTagger h;
     h = PartOfSpeech.trainCRF(train, xA, yA);
     System.out.println("CRF    Train Accuracy = "
     + StaticUtils.computeAccuracyS(h, train));
View Full Code Here
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
public class Newsgroups {

  public static void main(String[] args) throws IOException {
    // read in the data.
    ArrayList<ClassificationInstance> allData = (new NewsgroupsReader(
        new Alphabet(), new Alphabet())).readFile(args[0]);
    StaticUtils.shuffle(allData, 0);
    // randomly split data into training and testing part
    ArrayList<ClassificationInstance>[] tmp = StaticUtils.split(allData,
        (10));
    System.out.println(allData.get(0).xAlphabet.size());
    ArrayList<ClassificationInstance> train = tmp[0];
    ArrayList<ClassificationInstance> test = tmp[1];
    Alphabet xA = allData.get(0).xAlphabet;
    Alphabet yA = allData.get(0).yAlphabet;
    System.out.println("num Features = " + allData.get(0).xAlphabet.size());
    LinearClassifier h;
    h = trainMaxEnt(train, xA, yA);
    // print out accuracy
    System.out.println("MaxEnt Train Accuracy = "
View Full Code Here
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
      ResourceInstantiationException {

    // extract the document content as a string
    String text = document.getContent().toString();
    ArrayList<Element<Object>> elements = new ArrayList<Element<Object>>();
    Alphabet labelAlphabet = tagger.getYAlphabet();
    FeatureMap features = Factory.newFeatureMap();
    long start = 0;
    boolean wasSpace = true;
    ElementSequence<Element<Object>> sequence;
    SparseVector[] x;
    Object[] y;
    char[] chars;
    int[] labels;
    String chunk, label;
    long end, newStart, newEnd;
    int id;

    for (int i = 0; i < text.length(); i++) {

      if ((text.charAt(i) == ' ') || (text.charAt(i) == '\n')) {

        // if the previous character wasn't a space but this one is,
        // update the end offset and make a temporary token annotation
        if (!wasSpace) {

          end = i;
          id = outputAS.add(start, end, "TempToken", features);
          chars = document.getContent().getContent(start, end)
              .toString().toCharArray();

          for (Character c : chars) {

            // extract the chunk of the character
            chunk = c.toString();
            label = "?";

            // create an element with the chunk and label
            elements.add(new Element<Object>(chunk, label, outputAS
                .get(id)));
          }

          // create the x and y arrays
          x = new SparseVector[elements.size()];
          y = new Object[elements.size()];

          // fill in the labels
          for (int j = 0; j < y.length; j++)
            y[j] = "?";

          // create an element sequence, extract its features and
          // label it
          sequence = new ElementSequence(elements, xAlphabet,
              yAlphabet, x, y, document, inputAS);
          labels = tagger.label(extractor.process(sequence).x);
          elements = new ArrayList<Element<Object>>();

          if (labels.length == chars.length) {

            // update the new start offset
            newStart = start;

            for (int j = 1; j < chars.length - 1; j++) {

              // if a middle character is labeled "B",
              // update the new end offset, create a token
              // annotation
              // and update the new start offset
              if (labels[j] == labelAlphabet.lookupObject("B")) {

                newEnd = start + j;
                outputAS.add(newStart, newEnd, labelType,
                    features);
                newStart = newEnd;
View Full Code Here
63
64
65
66
67
68
69
70
71
    // load the pipeline and extractor
    pipeline = (ConditionalSerialAnalyserController)
    PersistenceManager.loadObjectFromUrl(path2xgapp);
    extractor = new GroovyPreprocessor(path2rules);
    sequences = new ArrayList<SequenceInstance>();
    xAlphabet = new Alphabet();
    yAlphabet = new Alphabet();
    this.path2xgapp = path2xgapp;
  }
View Full Code Here
71
72
73
74
75
76
77
78
            + "used to create the model");
    }
    extractor = (GroovyPreprocessor) in.readObject();
    tagger = (LinearTagger) in.readObject();
    in.close();
    xAlphabet = new Alphabet();
    yAlphabet = new Alphabet();
  }
View Full Code Here
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
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
    Factory.deleteResource(pipeline);

    // extract sentences from text
    AnnotationSet sentences = inputAS.get(sequenceType);
    ArrayList<Element<Object>> elements = new ArrayList<Element<Object>>();
    Alphabet labelAlphabet = tagger.getYAlphabet();
    OffsetComparator oc = new OffsetComparator();
    SparseVector[] x;
    Object[] y;
    int[] labels;
    ElementSequence<Element<Object>> sequence;
    Element<Object> element;
    ArrayList<Annotation> tokens;
    Iterator<Annotation> iterator;
    String chunk, label;

    for (Annotation sentence : sentences) {

      // extract tokens from sentence
      tokens = new ArrayList(inputAS.get(
          sentence.getStartNode().getOffset(),
          sentence.getEndNode().getOffset()).get(elementType));
      Collections.sort(tokens, oc);

      for (Annotation token : tokens) {

        // extract the chunk of the token
        chunk = document.getContent().getContent(
            token.getStartNode().getOffset(),
            token.getEndNode().getOffset()).toString();
        label = "?";

        // create an element with the chunk and label
        element = new Element<Object>(chunk, label, token);
        elements.add(element);
      }

      x = new SparseVector[elements.size()];
      y = new Object[elements.size()];

      // fill in the labels
      for (int i = 0; i < y.length; i++)
        y[i] = "?";

      // create an element sequence and extract the features
      sequence = new ElementSequence(elements, xAlphabet, yAlphabet, x,
          y, document, inputAS);
      elements = new ArrayList<Element<Object>>();

      // extract the labels for the tokens in the sentence
      labels = tagger.label(extractor.process(sequence).x);
      iterator = tokens.iterator();

      for (int i = 0; i < labels.length; i++) {

        // add the labels as new annotation features
        if (iterator.hasNext())
          iterator.next().getFeatures().put(labelType,
              labelAlphabet.lookupIndex(labels[i]));

        // shouldn't happen, but who knows?
        else
          System.err
          .println("Unexpected number of tokens in sentence!");
View Full Code Here
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
125
    Factory.deleteResource(pipeline);
   
    // extract sentences from text
    AnnotationSet sentences = inputAS.get(sequenceType);
    ArrayList<Element<Object>> elements = new ArrayList<Element<Object>>();
    Alphabet labelAlphabet = tagger.getYAlphabet();
    OffsetComparator oc = new OffsetComparator();
    SparseVector[] x;
    Object[] y;
    int[] labels;
    ElementSequence<Element<Object>> sequence;
    Element<Object> element;
    ArrayList<Annotation> tokens;
    String chunk, label;
   
    for (Annotation sentence : sentences) {
     
      // extract tokens from sentence
      tokens = new ArrayList(inputAS.get(sentence.getStartNode().getOffset(),
          sentence.getEndNode().getOffset()).get(elementType));
      Collections.sort(tokens, oc);
     
      for (Annotation token : tokens) {

        // extract the chunk of the token
        chunk = document.getContent().getContent(token.getStartNode().getOffset(),
            token.getEndNode().getOffset()).toString();
        label = "?";

        // create an element with the chunk and label
        element = new Element<Object>(chunk, label, token);
        elements.add(element);
      }
     
      x = new SparseVector[elements.size()];
      y = new Object[elements.size()];

      // fill in the labels
      for (int i = 0; i < y.length; i++)
        y[i] = "?";

      // create an element sequence and extract the features
      sequence = new ElementSequence(elements, xAlphabet, yAlphabet, x,
          y, document, inputAS);
      elements = new ArrayList<Element<Object>>();

      // extract the labels for the tokens in the sentence
      labels = tagger.label(extractor.process(sequence).x);
      FeatureMap features = Factory.newFeatureMap();
      long start, end;
     
      if (labels.length == tokens.size()) {
       
        for (int i = 0; i < tokens.size(); i++) {
         
          if (labels[i] == labelAlphabet.lookupObject("B-NP")) {
            start = tokens.get(i).getStartNode().getOffset();
            end = tokens.get(i).getEndNode().getOffset();
           
            while ((i + 1 < tokens.size()) &&
                (labels[i + 1] == labelAlphabet.lookupObject("I-NP"))) {
              end = tokens.get(i + 1).getEndNode().getOffset();
              i++;
            }
           
            outputAS.add(start, end, nChunkType, features);
          }
         
          else if (labels[i] == labelAlphabet.lookupObject("B-VP")) {
            start = tokens.get(i).getStartNode().getOffset();
            end = tokens.get(i).getEndNode().getOffset();
           
            while ((i + 1 < tokens.size()) &&
                (labels[i + 1] == labelAlphabet.lookupObject("I-VP"))) {
              end = tokens.get(i + 1).getEndNode().getOffset();
              i++;
            }
           
            outputAS.add(start, end, vChunkType, features);
View Full Code Here
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
125
126
127
128
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
    Factory.deleteResource(pipeline);

    // extract sentences from text
    AnnotationSet sentences = inputAS.get(sequenceType);
    ArrayList<Element<Object>> elements = new ArrayList<Element<Object>>();
    Alphabet labelAlphabet = tagger.getYAlphabet();
    OffsetComparator oc = new OffsetComparator();
    SparseVector[] x;
    Object[] y;
    int[] labels;
    ElementSequence<Element<Object>> sequence;
    Element<Object> element;
    ArrayList<Annotation> tokens;
    String chunk, label;

    for (Annotation sentence : sentences) {

      // extract tokens from sentence
      tokens = new ArrayList(inputAS.get(sentence.getStartNode().getOffset(),
          sentence.getEndNode().getOffset()).get(elementType));
      Collections.sort(tokens, oc);

      for (Annotation token : tokens) {

        // extract the chunk of the token
        chunk = document.getContent().getContent(token.getStartNode().getOffset(),
            token.getEndNode().getOffset()).toString();
        label = "?";

        // create an element with the chunk and label
        element = new Element<Object>(chunk, label, token);
        elements.add(element);
      }

      x = new SparseVector[elements.size()];
      y = new Object[elements.size()];

      // fill in the labels
      for (int i = 0; i < y.length; i++)
        y[i] = "?";

      // create an element sequence and extract the features
      sequence = new ElementSequence(elements, xAlphabet, yAlphabet, x,
          y, document, inputAS);
      elements = new ArrayList<Element<Object>>();

      // extract the labels for the tokens in the sentence
      labels = tagger.label(extractor.process(sequence).x);
      FeatureMap features = Factory.newFeatureMap();
      long start, end;

      if (labels.length == tokens.size()) {

        for (int i = 0; i < tokens.size(); i++) {

          if (labels[i] == labelAlphabet.lookupObject("B-DAT")) {
            start = tokens.get(i).getStartNode().getOffset();
            end = tokens.get(i).getEndNode().getOffset();

            while ((i + 1 < tokens.size()) &&
                (labels[i + 1] == labelAlphabet.lookupObject("I-DAT"))) {
              end = tokens.get(i + 1).getEndNode().getOffset();
              i++;
            }

            outputAS.add(start, end, datType, features);
          }

          else if (labels[i] == labelAlphabet.lookupObject("B-LOC")) {
            start = tokens.get(i).getStartNode().getOffset();
            end = tokens.get(i).getEndNode().getOffset();

            while ((i + 1 < tokens.size()) &&
                (labels[i + 1] == labelAlphabet.lookupObject("I-LOC"))) {
              end = tokens.get(i + 1).getEndNode().getOffset();
              i++;
            }

            outputAS.add(start, end, locType, features);
          }

          else if (labels[i] == labelAlphabet.lookupObject("B-ORG")) {
            start = tokens.get(i).getStartNode().getOffset();
            end = tokens.get(i).getEndNode().getOffset();

            while ((i + 1 < tokens.size()) &&
                (labels[i + 1] == labelAlphabet.lookupObject("I-ORG"))) {
              end = tokens.get(i + 1).getEndNode().getOffset();
              i++;
            }

            outputAS.add(start, end, orgType, features);
          }

          else if (labels[i] == labelAlphabet.lookupObject("B-PER")) {
            start = tokens.get(i).getStartNode().getOffset();
            end = tokens.get(i).getEndNode().getOffset();

            while ((i + 1 < tokens.size()) &&
                (labels[i + 1] == labelAlphabet.lookupObject("I-PER"))) {
              end = tokens.get(i + 1).getEndNode().getOffset();
              i++;
            }

            outputAS.add(start, end, perType, features);
View Full Code Here
TOP

Related Classes of types.Alphabet

  • classification.AdaBoost
  • com.ontotext.gate.edlin.Chunker
  • com.ontotext.gate.edlin.NER
  • com.ontotext.gate.edlin.SkeletonTagger
  • com.ontotext.gate.edlin.SkeletonTrainer
  • com.ontotext.gate.edlin.Splitter
  • com.ontotext.gate.edlin.Tokenizer
  • experiments.InternetAds
  • experiments.Newsgroups
  • experiments.PartOfSpeech
    • 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.