Examples of org.apache.mahout.common.RandomWrapper

org.apache.mahout.common.RandomWrapper

        }
    }


    @Test
    public void testRepeatedValues() {
        final RandomWrapper gen = RandomUtils.getRandom();


        // 5% of samples will be 0 or 1.0.  10% for each of the values 0.1 through 0.9
        AbstractContinousDistribution mix = new AbstractContinousDistribution() {
            @Override
            public double nextDouble() {
                return Math.rint(gen.nextDouble() * 10) / 10.0;
            }
        };


        TDigest dist = new TDigest((double) 1000);
        long t0 = System.nanoTime();

View Full Code Here

    }


    //@Test()
    // very slow running data generator
    public void testSizeControl() {
        RandomWrapper gen = RandomUtils.getRandom();
        System.out.printf("k\tsamples\tcompression\tsize1\tsize2\n");
        for (int k = 0; k < 40; k++) {
            for (int size : new int[]{10, 100, 1000, 10000}) {
                for (double compression : new double[]{2, 5, 10, 20, 50, 100, 200, 500, 1000}) {
                    TDigest dist = new TDigest(compression);
                    for (int i = 0; i < size * 1000; i++) {
                        dist.add(gen.nextDouble());
                    }
                    System.out.printf("%d\t%d\t%.0f\t%d\t%d\n", k, size, compression, dist.smallByteSize(), dist.byteSize());
                }
            }
        }

View Full Code Here

        System.out.printf("\n");
    }


    @Test
    public void testScaling() {
        RandomWrapper gen = RandomUtils.getRandom();


        System.out.printf("pass\tcompression\tq\terror\tsize\n");
        // change to 50 passes for better graphs
        for (int k = 0; k < 3; k++) {
            List<Double> data = Lists.newArrayList();
            for (int i = 0; i < 100000; i++) {
                data.add(gen.nextDouble());
            }
            Collections.sort(data);


            for (double compression : new double[]{2, 5, 10, 20, 50, 100, 200, 500, 1000}) {
                TDigest dist = new TDigest(compression);

View Full Code Here

        }
    }


    @Test
    public void testMerge() {
        RandomWrapper gen = RandomUtils.getRandom();


        for (int parts : new int[]{2, 5, 10, 20, 50, 100}) {
            List<Double> data = Lists.newArrayList();


            TDigest dist = new TDigest(100);
            dist.recordAllData();


            List<TDigest> many = Lists.newArrayList();
            for (int i = 0; i < 100; i++) {
                many.add(new TDigest(100).recordAllData());
            }


            // we accumulate the data into multiple sub-digests
            List<TDigest> subs = Lists.newArrayList();
            for (int i = 0; i < parts; i++) {
                subs.add(new TDigest(50).recordAllData());
            }


            for (int i = 0; i < 100000; i++) {
                double x = gen.nextDouble();
                data.add(x);
                dist.add(x);
                subs.get(i % parts).add(x);
            }
            dist.compress();

View Full Code Here

      // 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;
        }
      });
      this.weightMatrixList.add(weightMatrix);
      this.prevWeightUpdatesList.add(new DenseMatrix(row, sizePrevLayer));
      this.squashingFunctionList.add(squashingFunctionName);

View Full Code Here

      int sizePrevLayer = layerSizeList.get(layerIndex - 1);
      // 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;
        }
      });
      weightMatrixList.add(weightMatrix);
      prevWeightUpdatesList.add(new DenseMatrix(row, sizePrevLayer));
      squashingFunctionList.add(squashingFunctionName);

View Full Code Here


    this.numThreads = numThreads;
  }


  protected void initialize() throws TasteException {
    RandomWrapper random = RandomUtils.getRandom();
    userVectors = new double[dataModel.getNumUsers()][rank];
    itemVectors = new double[dataModel.getNumItems()][rank];


    double globalAverage = getAveragePreference();
    for (int userIndex = 0; userIndex < userVectors.length; userIndex++) {
      userVectors[userIndex][0] = globalAverage;
      userVectors[userIndex][USER_BIAS_INDEX] = 0; // will store user bias
      userVectors[userIndex][ITEM_BIAS_INDEX] = 1; // corresponding item feature contains item bias
      for (int feature = FEATURE_OFFSET; feature < rank; feature++) {
        userVectors[userIndex][feature] = random.nextGaussian() * NOISE;
      }
    }
    for (int itemIndex = 0; itemIndex < itemVectors.length; itemIndex++) {
      itemVectors[itemIndex][0] = 1; // corresponding user feature contains global average
      itemVectors[itemIndex][USER_BIAS_INDEX] = 1; // corresponding user feature contains user bias
      itemVectors[itemIndex][ITEM_BIAS_INDEX] = 0; // will store item bias
      for (int feature = FEATURE_OFFSET; feature < rank; feature++) {
        itemVectors[itemIndex][feature] = random.nextGaussian() * NOISE;
      }
    }
  }

View Full Code Here

0 1

TOP

Related Classes of org.apache.mahout.common.RandomWrapper

org.apache.commons.math3.random.MersenneTwister

org.apache.mahout.cf.taste.impl.recommender.svd.ParallelSGDFactorizer

org.apache.mahout.cf.taste.impl.recommender.svd.RatingSGDFactorizer

org.apache.mahout.classifier.mlp.NeuralNetwork

org.apache.mahout.ga.watchmaker.cd.hadoop.DatasetSplitTest

org.apache.mahout.math.stats.TDigestTest

org.uncommons.maths.random.MersenneTwisterRNG

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.