Package de.jungblut.math.sparse

Examples of de.jungblut.math.sparse.SparseDoubleVector

89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
  }

  @Override
  public DoubleVector predict(DoubleVector features) {
    if (features.isSparse()) {
      SparseDoubleVector tmp = new SparseDoubleVector(
          features.getDimension() + 1);
      tmp.set(0, 1d);
      Iterator<DoubleVectorElement> iterateNonZero = features.iterateNonZero();
      while (iterateNonZero.hasNext()) {
        DoubleVectorElement next = iterateNonZero.next();
        tmp.set(next.getIndex() + 1, next.getValue());
      }
      features = tmp;
    } else {
      features = new DenseDoubleVector(1d, features.toArray());
    }
View Full Code Here
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
     */
    @Override
    protected void reduce(Text key, Iterable<TextIntIntIntWritable> values,
        Context context) throws IOException, InterruptedException {

      SparseDoubleVector vector = new SparseDoubleVector(numTokens);
      for (TextIntIntIntWritable pair : values) {
        if (documentThreshold > pair.getSecond().get()) {
          double val = 0d;
          if (wordCount) {
            val = pair.getThird().get();
          } else {
            val = pair.getThird().get()
                * (FastMath.log(numDocs) - FastMath.log(pair.getSecond().get()));
          }
          vector.set(pair.getFourth().get(), val);
        }
      }

      context.write(key, new VectorWritable(vector));

View Full Code Here
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
      List<String[]> tokenizedDocuments, String[] dictionary) {

    List<DoubleVector> vectorList = new ArrayList<>(tokenizedDocuments.size());
    int oovIndex = Arrays.binarySearch(dictionary, OUT_OF_VOCABULARY);
    for (String[] arr : tokenizedDocuments) {
      DoubleVector vector = new SparseDoubleVector(dictionary.length);
      HashMultiset<String> set = HashMultiset.create(Arrays.asList(arr));
      for (String s : arr) {
        int foundIndex = Arrays.binarySearch(dictionary, s);
        // simply ignore tokens we don't know or that are spam
        if (foundIndex >= 0) {
          // the index is equal to its mapped dimension
          vector.set(foundIndex, set.count(s));
        } else if (oovIndex >= 0) {
          vector.set(oovIndex, 1);
        }
      }
      vectorList.add(vector);
    }
View Full Code Here
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
   */
  public static DoubleVector tfIdfVectorize(int numDocuments,
      String[] document, String[] dictionary, int[] termDocumentCount) {

    final int numTokens = dictionary.length;
    DoubleVector vector = new SparseDoubleVector(numTokens);
    HashMultiset<String> termFrequencySet = HashMultiset.create(Arrays
        .asList(document));

    int oovIndex = Arrays.binarySearch(dictionary, OUT_OF_VOCABULARY);
    double docLog = FastMath.log(numDocuments);

    for (String token : document) {
      int index = Arrays.binarySearch(dictionary, token);
      if (index >= 0) {
        double tfIdf = termFrequencySet.count(token)
            * (docLog - FastMath.log(termDocumentCount[index]));
        vector.set(index, tfIdf);
      } else {
        if (oovIndex >= 0) {
          vector.set(oovIndex, 1d);
        }
      }
    }
    return vector;
  }
View Full Code Here
114
115
116
117
118
119
120
121
122
123
124
      clz = 0;
    }
    if (binaryClassification) {
      return new DenseDoubleVector(new double[] { clz });
    } else {
      DoubleVector vec = outcomeDimension > 10 ? new SparseDoubleVector(
          outcomeDimension) : new DenseDoubleVector(outcomeDimension);
      vec.set(clz, 1);
      return vec;
    }
  }
View Full Code Here
109
110
111
112
113
114
115
116
117
118
    }
  }

  protected DoubleVector newInstance(DoubleVector v) {
    if (v.isSparse()) {
      return new SparseDoubleVector(v.getDimension());
    } else {
      return new DenseDoubleVector(v.getDimension());
    }
  }
View Full Code Here
109
110
111
112
113
114
115
116
117
118
119
120
121
    final int numRowIndices = in.readInt();
    for (int i = 0; i < numRowIndices; i++) {
      final int rowIndex = in.readInt();
      final int numColumns = in.readInt();
      DoubleVector row = new SparseDoubleVector(mat.getColumnCount());
      for (int j = 0; j < numColumns; j++) {
        row.set(in.readInt(), in.readDouble());
      }
      mat.setRowVector(rowIndex, row);
    }

    return mat;
View Full Code Here
140
141
142
143
144
145
146
147
148
149
150
    DoubleVector vector = null;
    switch (typeByte) {
      case SPARSE:
        int length = in.readInt();
        int dim = in.readInt();
        vector = new SparseDoubleVector(dim);
        for (int i = 0; i < length; i++) {
          int index = in.readInt();
          double value = in.readDouble();
          vector.set(index, value);
        }
View Full Code Here
18
19
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
        new CosineDistance());

    // we seperate stuff in two dimensions each
    DoubleVector left = new SingleEntryDoubleVector(0d);
    DoubleVector right = new SingleEntryDoubleVector(1d);
    DoubleVector v1 = new SparseDoubleVector(4);
    v1.set(0, 1d);
    v1.set(1, 1d);

    DoubleVector v2 = new SparseDoubleVector(4);
    v2.set(2, 1d);
    v2.set(3, 2.5);

    DoubleVector v3 = new SparseDoubleVector(4);
    v3.set(0, 2d);
    v3.set(1, 2d);

    DoubleVector v4 = new SparseDoubleVector(4);
    v4.set(2, 0.5);
    v4.set(3, 1.5);

    DoubleVector[] trainingSet = new DoubleVector[] { v1, v2, v3, v4 };
    DoubleVector[] outcomeSet = new DoubleVector[] { left, right, left, right };

    neighbours.train(trainingSet, outcomeSet);

    DoubleVector predict = neighbours.predict(v4);
    assertEquals(right, predict);

    predict = neighbours.predict(v2);
    assertEquals(right, predict);

    predict = neighbours.predict(v1);
    assertEquals(left, predict);

    predict = neighbours.predict(v3);
    assertEquals(left, predict);

    // predict between, slightly to the right
    DoubleVector vx = new SparseDoubleVector(4);
    vx.set(1, 1d);
    vx.set(3, 2.5);

    predict = neighbours.predict(vx);
    assertEquals(right, predict);
  }
View Full Code Here
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
  @Test
  public void testVectorInvertedIndex() {

    InvertedIndex<DoubleVector, Integer> invIndex = InvertedIndex
        .createVectorIndex(new CosineDistance());
    DoubleVector v1 = new SparseDoubleVector(4);
    v1.set(1, 0.6931471805599453);
    v1.set(0, 1.3862943611198906);
    DoubleVector v2 = new SparseDoubleVector(4);
    v2.set(2, 0.6931471805599453);
    v2.set(1, 1.3862943611198906);

    invIndex.build(Arrays.asList(v1, v2));

    DoubleVector v3 = new SparseDoubleVector(4);
    v3.set(3, 0.2);
    v3.set(1, 1);
    List<DistanceResult<DoubleVector>> res = invIndex.query(v3);
    assertEquals(2, res.size());
    assertEquals(v2, res.get(0).get());
    assertEquals(v1, res.get(1).get());
View Full Code Here
TOP

Related Classes of de.jungblut.math.sparse.SparseDoubleVector

  • de.jungblut.classification.bayes.MultinomialNaiveBayesTest
  • de.jungblut.classification.eval.EvaluationSplitTest
  • de.jungblut.classification.knn.SparseKNearestNeighboursTest
  • de.jungblut.classification.regression.LogisticRegression
  • de.jungblut.classification.tree.DecisionTree
  • de.jungblut.datastructure.InvertedIndexTest
  • de.jungblut.datastructure.KDTreeTest
  • de.jungblut.math.activation.AbstractActivationFunction
  • de.jungblut.ner.SparseFeatureExtractorHelper
  • de.jungblut.nlp.MinHashTest
    • 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.