Package org.apache.mahout.math

Examples of org.apache.mahout.math.Vector.iterateNonZero()

52
53
54
55
56
57
58
59
60
61
62
    if (v1.size() != v2.size()) {
      throw new CardinalityException(v1.size(), v2.size());
    }
    double result = 0.0;
    Vector vector = v1.minus(v2);
    Iterator<Vector.Element> iter = vector.iterateNonZero();
    while (iter.hasNext()) {
      Vector.Element e = iter.next();
      double d = Math.abs(e.get());
      if (d > result) {
        result = d;
View Full Code Here
143
144
145
146
147
148
149
150
151
152
153
    for (PrefAndSimilarityColumnWritable prefAndSimilarityColumn : values) {
      Vector simColumn = prefAndSimilarityColumn.getSimilarityColumn();
      float prefValue = prefAndSimilarityColumn.getPrefValue();
      /* count the number of items used for each prediction */
      Iterator<Vector.Element> usedItemsIterator = simColumn.iterateNonZero();
      while (usedItemsIterator.hasNext()) {
        int itemIDIndex = usedItemsIterator.next().index();
        numberOfSimilarItemsUsed.setQuick(itemIDIndex, numberOfSimilarItemsUsed.getQuick(itemIDIndex) + 1);
      }

View Full Code Here
82
83
84
85
86
87
88
89
90
91
92
   */
  @Override
  public double distance(Vector v1, Vector v2) {
    Vector distVector = v1.minus(v2);
    double sum = 0.0;
    Iterator<Element> it = distVector.iterateNonZero();
    while (it.hasNext()) {
      Element e = it.next();
      sum += Math.pow(Math.abs(e.get()), exponent);
    }
    return Math.pow(sum, 1.0 / exponent);
View Full Code Here
69
70
71
72
73
74
75
76
77
78
79
    ctx.getCounter(Elements.NEGLECTED).increment(numElementsBeforePruning - numElementsAfterPruning);

    DistributedRowMatrix.MatrixEntryWritable entry = new DistributedRowMatrix.MatrixEntryWritable();
    int colIndex = TasteHadoopUtils.idToIndex(rowIndex.get());
    entry.setCol(colIndex);
    Iterator<Vector.Element> iterator = vector.iterateNonZero();
    while (iterator.hasNext()) {
      Vector.Element elem = iterator.next();
      entry.setRow(elem.index());
      entry.setVal(elem.get());
      ctx.write(new IntWritable(elem.index()), entry);
View Full Code Here
49
50
51
52
53
54
55
56
57
58
59
  @Override
  protected void map(IntWritable key, VectorWritable value, Context context) throws IOException, InterruptedException {
    Vector vector = value.get();
    int label = key.get();
    double sigmaK = labelSum.get(label);
    Iterator<Element> it = vector.iterateNonZero();
    while (it.hasNext()) {
      Element e = it.next();
      double numerator = featureSum.get(e.index()) - e.get() + alphaI;
      double denominator = totalSum - sigmaK + vocabCount;
      double weight = Math.log(numerator / denominator);
View Full Code Here
130
131
132
133
134
135
136
137
138
139
140
      boolean firstIsOutFrag =  ((VectorWritable)v.get(0)).get().size() == outCardinality;
      Vector outFrag = firstIsOutFrag ? ((VectorWritable)v.get(0)).get() : ((VectorWritable)v.get(1)).get();
      Vector multiplier = firstIsOutFrag ? ((VectorWritable)v.get(1)).get() : ((VectorWritable)v.get(0)).get();

      VectorWritable outVector = new VectorWritable();
      Iterator<Vector.Element> it = multiplier.iterateNonZero();
      while (it.hasNext()) {
        Vector.Element e = it.next();
        row.set(e.index());
        outVector.set(outFrag.times(e.get()));
        out.collect(row, outVector);
View Full Code Here
56
57
58
59
60
61
62
63
64
65
66
    double[] logTotals = new double[state.getNumTopics()];
    Arrays.fill(logTotals, Double.NEGATIVE_INFINITY);
   
    // Output sufficient statistics for each word. == pseudo-log counts.
    DoubleWritable v = new DoubleWritable();
    for (Iterator<Vector.Element> iter = wordCounts.iterateNonZero(); iter.hasNext();) {
      Vector.Element e = iter.next();
      int w = e.index();
     
      for (int k = 0; k < state.getNumTopics(); ++k) {
        v.set(doc.phi(k, w) + Math.log(e.get()));
View Full Code Here
362
363
364
365
366
367
368
369
370
371
372
         "This is probably because the --numWords argument is set too small.  \n"
         + "\tIt needs to be >= than the number of words (terms actually) in the corpus and can be \n"
         + "\tlarger if some storage inefficiency can be tolerated.", e1);
      }

      for (Iterator<Vector.Element> iter = wordCounts.iterateNonZero(); iter.hasNext();) {
        Vector.Element e = iter.next();
        int w = e.index();

        for (int k = 0; k < state.getNumTopics(); ++k) {
          double vwUpdate = doc.phi(k, w) + Math.log(e.get());
View Full Code Here
64
65
66
67
68
69
70
71
72
73
74
        // classification yields probabilities
        Vector probabilities = classifier.classify(vector);
        // policy selects weights for models given those probabilities
        Vector weights = policy.select(probabilities);
        // training causes all models to observe data
        for (Iterator<Vector.Element> it = weights.iterateNonZero(); it.hasNext();) {
          int index = it.next().index();
          classifier.train(index, vector, weights.get(index));
        }
      }
      // compute the posterior models
View Full Code Here
103
104
105
106
107
108
109
110
111
112
113
        // classification yields probabilities
        Vector probabilities = classifier.classify(vector);
        // policy selects weights for models given those probabilities
        Vector weights = policy.select(probabilities);
        // training causes all models to observe data
        for (Iterator<Vector.Element> it = weights.iterateNonZero(); it
            .hasNext();) {
          int index = it.next().index();
          classifier.train(index, vector, weights.get(index));
        }
      }
View Full Code Here
TOP
Copyright © 2018 www.massapi.com. 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.