Examples of aima.core.probability.util.ProbabilityTable

aima.core.probability.util.ProbabilityTable
A Utility Class for associating values with a set of finite Random Variables. This is also the default implementation of the CategoricalDistribution and Factor interfaces (as they are essentially dependent on the same underlying data structures). @author Ciaran O'Reilly

          "The number of parent value arguments ["
              + parentValues.length
              + "] is not equal to the number of parents ["
              + parents.size() + "] for this CPT.");
    }
    final ProbabilityTable cc = new ProbabilityTable(getOn());
    ProbabilityTable.Iterator pti = new ProbabilityTable.Iterator() {
      private int idx = 0;


      @Override
      public void iterate(Map<RandomVariable, Object> possibleAssignment,
          double probability) {
        cc.getValues()[idx] = probability;
        idx++;
      }


      @Override
      public Object getPostIterateValue() {

View Full Code Here

    Set<RandomVariable> fofVars = new LinkedHashSet<RandomVariable>(
        table.getFor());
    for (AssignmentProposition ap : evidence) {
      fofVars.remove(ap.getTermVariable());
    }
    final ProbabilityTable fof = new ProbabilityTable(fofVars);
    // Otherwise need to iterate through the table for the
    // non evidence variables.
    final Object[] termValues = new Object[fofVars.size()];
    ProbabilityTable.Iterator di = new ProbabilityTable.Iterator() {
      public void iterate(Map<RandomVariable, Object> possibleWorld,
          double probability) {
        if (0 == termValues.length) {
          fof.getValues()[0] += probability;
        } else {
          int i = 0;
          for (RandomVariable rv : fof.getFor()) {
            termValues[i] = possibleWorld.get(rv);
            i++;
          }
          fof.getValues()[fof.getIndex(termValues)] += probability;
        }
      }


      public Object getPostIterateValue() {
        return null; // N/A

View Full Code Here

  public void test_divideBy() {
    RandomVariable xRV = new RandVar("X", new BooleanDomain());
    RandomVariable yRV = new RandVar("Y", new BooleanDomain());
    RandomVariable zRV = new RandVar("Z", new BooleanDomain());


    ProbabilityTable xyzD = new ProbabilityTable(new double[] {
        // X = true, Y = true, Z = true
        1.0,
        // X = true, Y = true, Z = false
        2.0,
        // X = true, Y = false, Z = true
        3.0,
        // X = true, Y = false, Z = false
        4.0,
        // X = false, Y = true, Z = true
        5.0,
        // X = false, Y = true, Z = false
        6.0,
        // X = false, Y = false, Z = true
        7.0,
        // X = false, Y = false, Z = false
        8.0, }, xRV, yRV, zRV);
    ProbabilityTable xzyD = new ProbabilityTable(new double[] {
        // X = true, Z = true, Y = true
        1.0,
        // X = true, Z = true, Y = false
        3.0,
        // X = true, Z = false, Y = true
        2.0,
        // X = true, Z = false, Y = false
        4.0,
        // X = false, Z = true, Y = true
        5.0,
        // X = false, Z = true, Y = false
        7.0,
        // X = false, Z = false, Y = true
        6.0,
        // X = false, Z = false, Y = false
        8.0, }, xRV, zRV, yRV);
    ProbabilityTable zxyD = new ProbabilityTable(new double[] {
        // Z = true, X = true, Y = true
        1.0,
        // Z = true, X = true, Y = false
        3.0,
        // Z = true, X = false, Y = true
        5.0,
        // Z = true, X = false, Y = false
        7.0,
        // Z = false, X = true, Y = true
        2.0,
        // Z = false, X = true, Y = false
        4.0,
        // Z = false, X = false, Y = true
        6.0,
        // Z = false, X = false, Y = false
        8.0, }, zRV, xRV, yRV);
    ProbabilityTable zD = new ProbabilityTable(new double[] { 0.5, 0.2 },
        zRV);
    // The identity distribution (to order results for comparison purposes)
    ProbabilityTable iD = new ProbabilityTable(new double[] { 1.0 });
    // Ensure the order of the dividends
    // makes no difference to the result
    Assert.assertArrayEquals(xyzD.divideBy(zD).getValues(),
        xzyD.divideBy(zD).pointwiseProductPOS(iD, xRV, yRV, zRV)
            .getValues(), DELTA_THRESHOLD);

View Full Code Here

  public void test_pointwiseProduct() {
    RandomVariable xRV = new RandVar("X", new BooleanDomain());
    RandomVariable yRV = new RandVar("Y", new BooleanDomain());
    RandomVariable zRV = new RandVar("Z", new BooleanDomain());


    ProbabilityTable xyD = new ProbabilityTable(new double[] {
        // X = true, Y = true
        1.0,
        // X = true, Y = false
        2.0,
        // X = false, Y = true
        3.0,
        // X = false, Y = false
        4.0 }, xRV, yRV);
    ProbabilityTable zD = new ProbabilityTable(new double[] { 3.0, 7.0 },
        zRV);


    // Not commutative
    Assert.assertArrayEquals(new double[] { 3.0, 7.0, 6.0, 14.0, 9.0, 21.0,
        12.0, 28.0 }, xyD.pointwiseProduct(zD).getValues(),
        DELTA_THRESHOLD);
    Assert.assertArrayEquals(new double[] { 3.0, 6.0, 9.0, 12.0, 7.0, 14.0,
        21.0, 28.0 }, zD.pointwiseProduct(xyD).getValues(),
        DELTA_THRESHOLD);
  }

View Full Code Here

  public void test_pointwiseProductPOS() {
    RandomVariable xRV = new RandVar("X", new BooleanDomain());
    RandomVariable yRV = new RandVar("Y", new BooleanDomain());
    RandomVariable zRV = new RandVar("Z", new BooleanDomain());


    ProbabilityTable xyD = new ProbabilityTable(new double[] {
        // X = true, Y = true
        1.0,
        // X = true, Y = false
        2.0,
        // X = false, Y = true
        3.0,
        // X = false, Y = false
        4.0 }, xRV, yRV);
    ProbabilityTable zD = new ProbabilityTable(new double[] { 3.0, 7.0 },
        zRV);


    // Make commutative by specifying an order for the product
    Assert.assertArrayEquals(xyD.pointwiseProduct(zD).getValues(), zD
        .pointwiseProductPOS(xyD, xRV, yRV, zRV).getValues(),
        DELTA_THRESHOLD);
  }

View Full Code Here

  @Test
  public void test_iterateOverTable_fixedValues() {
    RandVar aRV = new RandVar("A", new BooleanDomain());
    RandVar bRV = new RandVar("B", new BooleanDomain());
    RandVar cRV = new RandVar("C", new BooleanDomain());
    ProbabilityTable ptABC = new ProbabilityTable(new double[] {
        // A = true, B = true, C = true
        1.0,
        // A = true, B = true, C = false
        10.0,
        // A = true, B = false, C = true
        100.0,
        // A = true, B = false, C = false
        1000.0,
        // A = false, B = true, C = true
        10000.0,
        // A = false, B = true, C = false
        100000.0,
        // A = false, B = false, C = true
        1000000.0,
        // A = false, B = false, C = false
        10000000.0 }, aRV, bRV, cRV);


    final List<Double> answer = new ArrayList<Double>();
    ProbabilityTable.Iterator pti = new ProbabilityTable.Iterator() {


      @Override
      public void iterate(Map<RandomVariable, Object> possibleAssignment,
          double probability) {
        answer.add(probability);
      }


      @Override
      public Object getPostIterateValue() {
        return null; // N/A
      }
    };


    answer.clear();
    ptABC.iterateOverTable(pti, new AssignmentProposition(aRV, true));
    Assert.assertEquals(1111.0, sumOf(answer), DELTA_THRESHOLD);


    answer.clear();
    ptABC.iterateOverTable(pti, new AssignmentProposition(aRV, false));
    Assert.assertEquals(11110000.0, sumOf(answer), DELTA_THRESHOLD);


    answer.clear();
    ptABC.iterateOverTable(pti, new AssignmentProposition(bRV, true));
    Assert.assertEquals(110011.0, sumOf(answer), DELTA_THRESHOLD);


    answer.clear();
    ptABC.iterateOverTable(pti, new AssignmentProposition(bRV, false));
    Assert.assertEquals(11001100.0, sumOf(answer), DELTA_THRESHOLD);


    answer.clear();
    ptABC.iterateOverTable(pti, new AssignmentProposition(cRV, true));
    Assert.assertEquals(1010101.0, sumOf(answer), DELTA_THRESHOLD);


    answer.clear();
    ptABC.iterateOverTable(pti, new AssignmentProposition(cRV, false));
    Assert.assertEquals(10101010.0, sumOf(answer), DELTA_THRESHOLD);


    answer.clear();
    ptABC.iterateOverTable(pti, new AssignmentProposition(bRV, true),
        new AssignmentProposition(cRV, true));
    Assert.assertEquals(10001.0, sumOf(answer), DELTA_THRESHOLD);
  }

View Full Code Here

0 1 2

TOP

Related Classes of aima.core.probability.util.ProbabilityTable

aima.core.probability.bayes.approx.GibbsAsk

aima.core.probability.bayes.approx.LikelihoodWeighting

aima.core.probability.bayes.approx.RejectionSampling

aima.core.probability.bayes.exact.EnumerationAsk

aima.core.probability.bayes.impl.CPT

aima.core.probability.bayes.model.FiniteBayesModel

aima.core.probability.full.FullJointDistributionModel

aima.core.probability.hmm.impl.HMM

aima.core.probability.temporal.generic.ForwardBackward

aima.core.util.math.MixedRadixNumber

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.