Package de.jungblut.math

Examples of de.jungblut.math.DoubleVector

29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
  }

  @Override
  public double calculateError(DoubleVector y, DoubleVector hypothesis) {

    DoubleVector negativeOutcome = y.subtractFrom(1.0d);
    DoubleVector inverseOutcome = y.multiply(-1d);
    DoubleVector negativeHypo = hypothesis.subtractFrom(1d);
    DoubleVector negativeLogHypo = MathUtils.logVector(negativeHypo);
    DoubleVector positiveLogHypo = MathUtils.logVector(hypothesis);
    DoubleVector negativePenalty = negativeOutcome.multiply(negativeLogHypo);
    DoubleVector positivePenalty = inverseOutcome.multiply(positiveLogHypo);

    return (positivePenalty.subtract(negativePenalty)).sum();
  }
View Full Code Here
44
45
46
47
48
49
50
51
52
53
54
55
56
    if (matrix.isSparse()) {
      // if we have a sparse matrix, it is more efficient to loop over the
      // sparse row vectors
      int[] rows = matrix.rowIndices();
      for (int row : rows) {
        DoubleVector rowVector = matrix.getRowVector(row);
        if (rowVector.getLength() > 0) {
          DoubleVector apply = apply(rowVector);
          newInstance.setRowVector(row, apply);
        }
      }
    } else {
      // on dense matrices we can be faster by directly looping over the items
View Full Code Here
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
    return newInstance;
  }

  @Override
  public DoubleVector gradient(DoubleVector vector) {
    DoubleVector newInstance = newInstance(vector);
    if (vector.isSparse()) {
      Iterator<DoubleVectorElement> iterateNonZero = vector.iterateNonZero();
      while (iterateNonZero.hasNext()) {
        DoubleVectorElement next = iterateNonZero.next();
        newInstance.set(next.getIndex(), gradient(next.getValue()));
      }
    } else {
      for (int i = 0; i < vector.getDimension(); i++) {
        newInstance.set(i, gradient(vector.get(i)));
      }
    }
    return newInstance;
  }
View Full Code Here
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
   */
  public static DoubleVector readImageAsGreyScale(BufferedImage img) {
    final int h = img.getHeight();
    final int w = img.getWidth();

    DoubleVector vector = new DenseDoubleVector(h * w);
    int[] rgb = img.getRGB(0, 0, w, h, null, 0, w);
    for (int i = 0; i < rgb.length; i++) {
      int red = (rgb[i] >> 16) & 0xFF;
      int green = (rgb[i] >> 8) & 0xFF;
      int blue = (rgb[i] >> 0) & 0xFF;
      vector.set(i, (red + green + blue) / 3d);
    }
    return vector;
  }
View Full Code Here
30
31
32
33
34
35
36
37
38
39
40
41
  @Override
  public double calculateError(DoubleMatrix y, DoubleMatrix hypothesis) {
    double hammingSum = 0d;
    // we now loop row-wise over the matrices
    for (int row = 0; row < y.getRowCount(); row++) {
      DoubleVector yRow = y.getRowVector(row);
      DoubleVector hypRow = hypothesis.getRowVector(row);
      hammingSum += calculateError(yRow, hypRow);
    }

    return hammingSum / y.getRowCount();
  }
View Full Code Here
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
  @Override
  protected void reduce(ClusterCenter key, Iterable<VectorWritable> values,
      Context context) throws IOException, InterruptedException {

    List<VectorWritable> vectorList = new ArrayList<>();
    DoubleVector newCenter = null;
    for (VectorWritable value : values) {
      vectorList.add(new VectorWritable(value));
      if (newCenter == null)
        newCenter = value.getVector().deepCopy();
      else
        newCenter = newCenter.add(value.getVector());
    }

    newCenter = newCenter.divide(vectorList.size());
    ClusterCenter center = new ClusterCenter(newCenter);
    centers.add(center);
    for (VectorWritable vector : vectorList) {
      context.write(center, vector);
    }
View Full Code Here
173
174
175
176
177
178
179
180
181
182
183
      // in every iteration we initialize a new model that is a copy of the old
      DoubleMatrix transitionProbabilityMatrix = this.transitionProbabilityMatrix
          .deepCopy();
      DoubleMatrix emissionProbabilityMatrix = this.emissionProbabilityMatrix
          .deepCopy();
      DoubleVector hiddenPriorProbability = this.hiddenPriorProbability
          .deepCopy();

      // expectation step
      alpha = forward(alpha, transitionProbabilityMatrix,
          emissionProbabilityMatrix, hiddenPriorProbability, features);
View Full Code Here
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
    for (int row = 0; row < transitionProbabilityMatrix.getRowCount(); row++) {
      // note that we are using row vectors here, because dense matrices give us
      // the underlying array wrapped by the vector object so we can directly
      // mutate the values beneath
      DoubleVector rowVector = transitionProbabilityMatrix.getRowVector(row);
      rowVector = rowVector.divide(rowVector.sum());
      if (log) {
        rowVector = rowVector.log();
      }
      transitionProbabilityMatrix.setRowVector(row, rowVector);
      rowVector = emissionProbabilitiyMatrix.getRowVector(row);
      rowVector = rowVector.divide(rowVector.sum());
      if (log) {
        rowVector = rowVector.log();
      }
      emissionProbabilitiyMatrix.setRowVector(row, rowVector);
    }
  }
View Full Code Here
96
97
98
99
100
101
102
103
104
105
106
    DoubleMatrix thetaGradient = positiveAssociations.subtract(
        negativeAssociations).divide(data.getRowCount());

    // calculate the weight decay and apply it
    if (lambda != 0d) {
      DoubleVector bias = thetaGradient.getColumnVector(0);
      thetaGradient = thetaGradient.subtract(thetaGradient.multiply(lambda
          / data.getRowCount()));
      thetaGradient.setColumnVector(0, bias);
    }
View Full Code Here
59
60
61
62
63
64
65
66
67
68
69
    }
  }

  @Override
  public DoubleVector predictProbability(DoubleVector features) {
    DoubleVector prediction = predict(features);
    if (numOutcomes != 2) {
      prediction = prediction.divide(prediction.sum());
    }
    return prediction;
  }
View Full Code Here
TOP

Related Classes of de.jungblut.math.DoubleVector

  • de.jungblut.classification.AbstractPredictor
  • de.jungblut.classification.bayes.MultinomialNaiveBayes
  • de.jungblut.classification.bayes.MultinomialNaiveBayesTest
  • de.jungblut.classification.eval.Evaluator
  • de.jungblut.classification.knn.AbstractKNearestNeighbours
  • de.jungblut.classification.knn.KNearestNeighboursTest
  • de.jungblut.classification.knn.SparseKNearestNeighboursTest
  • de.jungblut.classification.meta.Voter
  • de.jungblut.classification.nn.MultilayerPerceptron
  • de.jungblut.classification.nn.MultilayerPerceptronCostFunction
    • 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.