Package org.apache.mahout.math.function

Examples of org.apache.mahout.math.function.DoubleFunction

79
80
81
82
83
84
85
86
87
88
89
  @Test
  public void testRandomStructuredModelViaMR() throws Exception {
    int numGeneratingTopics = 3;
    int numTerms = 9;
    Matrix matrix = ClusteringTestUtils.randomStructuredModel(numGeneratingTopics, numTerms, new DoubleFunction() {
      @Override
      public double apply(double d) {
        return 1.0 / Math.pow(d + 1.0, 3);
      }
    });
View Full Code Here
13
14
15
16
17
18
19
20
21
22
23
        check(1000, getDistribution(0.5), new ForeignKeySampler(1000, 0.5));
        check(1000, getDistribution(1), new ForeignKeySampler(1000, 0.3));
    }

    private DoubleFunction getDistribution(final double alpha) {
        return new DoubleFunction() {
            @Override
            public double apply(double rank) {
                return Math.pow(rank + 1, -alpha);
            }
        };
View Full Code Here
282
283
284
285
286
287
288
289
290
291
292
      // Row count equals to size of current size and column count equal to size of previous layer
      int row = isFinalLayer ? actualSize : actualSize - 1;
      Matrix weightMatrix = new DenseMatrix(row, sizePrevLayer);
      // Initialize weights
      final RandomWrapper rnd = RandomUtils.getRandom();
      weightMatrix.assign(new DoubleFunction() {
        @Override
        public double apply(double value) {
          return rnd.nextDouble() - 0.5;
        }
      });
View Full Code Here
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
    List<Vector> internalResults = getOutputInternal(inputInstance);

    Vector deltaVec = new DenseVector(layerSizeList.get(layerSizeList.size() - 1));
    Vector output = internalResults.get(internalResults.size() - 1);

    final DoubleFunction derivativeSquashingFunction =
        NeuralNetworkFunctions.getDerivativeDoubleFunction(squashingFunctionList.get(squashingFunctionList.size() - 1));

    final DoubleDoubleFunction costFunction =
        NeuralNetworkFunctions.getDerivativeDoubleDoubleFunction(costFunctionName);

    Matrix lastWeightMatrix = weightMatrixList.get(weightMatrixList.size() - 1);

    for (int i = 0; i < deltaVec.size(); ++i) {
      double costFuncDerivative = costFunction.apply(labels.get(i), output.get(i + 1));
      // Add regularization
      costFuncDerivative += regularizationWeight * lastWeightMatrix.viewRow(i).zSum();
      deltaVec.set(i, costFuncDerivative);
      deltaVec.set(i, deltaVec.get(i) * derivativeSquashingFunction.apply(output.get(i + 1)));
    }

    // Start from previous layer of output layer
    for (int layer = layerSizeList.size() - 2; layer >= 0; --layer) {
      deltaVec = backPropagate(layer, deltaVec, internalResults, weightUpdateMatrices[layer]);
View Full Code Here
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
   */
  private Vector backPropagate(int currentLayerIndex, Vector nextLayerDelta,
                               List<Vector> outputCache, Matrix weightUpdateMatrix) {

    // Get layer related information
    final DoubleFunction derivativeSquashingFunction =
        NeuralNetworkFunctions.getDerivativeDoubleFunction(squashingFunctionList.get(currentLayerIndex));
    Vector curLayerOutput = outputCache.get(currentLayerIndex);
    Matrix weightMatrix = weightMatrixList.get(currentLayerIndex);
    Matrix prevWeightMatrix = prevWeightUpdatesList.get(currentLayerIndex);

    // Next layer is not output layer, remove the delta of bias neuron
    if (currentLayerIndex != layerSizeList.size() - 2) {
      nextLayerDelta = nextLayerDelta.viewPart(1, nextLayerDelta.size() - 1);
    }

    Vector delta = weightMatrix.transpose().times(nextLayerDelta);

    delta = delta.assign(curLayerOutput, new DoubleDoubleFunction() {
      @Override
      public double apply(double deltaElem, double curLayerOutputElem) {
        return deltaElem * derivativeSquashingFunction.apply(curLayerOutputElem);
      }
    });

    // Update weights
    for (int i = 0; i < weightUpdateMatrix.rowSize(); ++i) {
View Full Code Here
156
157
158
159
160
161
162
163
164
165
166
   * Uniform [0,1) matrix generator function
   *
   * @param seed generator seed
   */
  public static final IntIntFunction uniformGenerator(final int seed) {
    return Functions.chain(new DoubleFunction() {
      @Override
      public double apply(double x) {
        return (x + 1.0) / 2.0;
      }
    }, uniformSymmetricGenerator(seed));
View Full Code Here
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
      // finish off pending regularization
      model.close();

      Matrix beta = model.getBeta();
      maxBeta = beta.aggregate(Functions.MAX, Functions.ABS);
      nonZeros = beta.aggregate(Functions.PLUS, new DoubleFunction() {
        @Override
        public double apply(double v) {
          return Math.abs(v) > 1.0e-6 ? 1 : 0;
        }
      });
      positive = beta.aggregate(Functions.PLUS, new DoubleFunction() {
        @Override
        public double apply(double v) {
          return v > 0 ? 1 : 0;
        }
      });
View Full Code Here
76
77
78
79
80
81
82
83
84
85
86
        DoubleMatrix2D UT = new DenseDoubleMatrix2D(m, k);
        double [] work = new double [U.columns()];

        // Colt functions
        DoubleDoubleFunction invDiv = Functions.swapArgs(Functions.DIV);
        DoubleFunction plusEps = Functions.plus(eps);

        if (stopThreshold >= 0)
        {
            updateApproximationError();
        }
View Full Code Here
76
77
78
79
80
81
82
83
84
85
86
        double [] work = new double [U.columns()];

        // Colt functions
        DoubleDoubleFunction invDiv = Functions.swapArgs(Functions.DIV);
        DoubleDoubleFunction sqrtMult = Functions.chain(Functions.SQRT, Functions.MULT);
        DoubleFunction plusEps = Functions.plus(eps);

        if (stopThreshold >= 0)
        {
            updateApproximationError();
        }
View Full Code Here
65
66
67
68
69
70
71
72
73
74
75
        DoubleMatrix2D T = new DenseDoubleMatrix2D(k, k);
        DoubleMatrix2D UT1 = new DenseDoubleMatrix2D(A.rows(), k);
        DoubleMatrix2D UT2 = new DenseDoubleMatrix2D(A.rows(), k);
        DoubleMatrix2D VT1 = new DenseDoubleMatrix2D(A.columns(), k);
        DoubleMatrix2D VT2 = new DenseDoubleMatrix2D(A.columns(), k);
        DoubleFunction plusEps = Functions.plus(eps);

        if (stopThreshold >= 0)
        {
            updateApproximationError();
        }
View Full Code Here
TOP

Related Classes of org.apache.mahout.math.function.DoubleFunction

  • com.mapr.stats.bandit.ContextualBayesBanditTest
  • com.mapr.synth.ForeignKeySamplerTest
  • mia.classifier.ch16.train.TrainNewsGroups
  • org.apache.mahout.cf.taste.impl.recommender.svd.ParallelSGDFactorizerTest
  • org.apache.mahout.classifier.mlp.NeuralNetwork
  • org.apache.mahout.classifier.sgd.AbstractOnlineLogisticRegression
  • org.apache.mahout.classifier.sgd.ModelSerializerTest
  • org.apache.mahout.classifier.sgd.SGDHelper
  • org.apache.mahout.classifier.sgd.TrainNewsGroups
  • org.apache.mahout.clustering.ClusteringTestUtils
    • 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.