/// <include file='FactorDocs.xml' path='factor_docs/message_op_class[@name="WrappedGaussianProductOp"]/message_doc[@name="AAverageConditional(WrappedGaussian, double, WrappedGaussian)"]/*'/> public static WrappedGaussian AAverageConditional([SkipIfUniform] WrappedGaussian Product, double B, WrappedGaussian result) { result.Period = Product.Period / B; result.Gaussian = GaussianProductOp.AAverageConditional(Product.Gaussian, B); result.Normalize(); return(result); }
private void GaussianProductOp_APointMass(double aMean, Gaussian Product, Gaussian B) { bool isProper = Product.IsProper(); Gaussian A = Gaussian.PointMass(aMean); Gaussian result = GaussianProductOp.AAverageConditional(Product, A, B); Console.WriteLine("{0}: {1}", A, result); Gaussian result2 = isProper ? GaussianProductOp_Slow.AAverageConditional(Product, A, B) : result; Console.WriteLine("{0}: {1}", A, result2); Assert.True(result.MaxDiff(result2) < 1e-6); var Amsg = InnerProductOp_PointB.BAverageConditional(Product, DenseVector.FromArray(B.GetMean()), new PositiveDefiniteMatrix(new double[, ] { { B.GetVariance() } }), VectorGaussian.PointMass(aMean), VectorGaussian.Uniform(1)); //Console.WriteLine("{0}: {1}", A, Amsg); Assert.True(result.MaxDiff(Amsg.GetMarginal(0)) < 1e-6); double prevDiff = double.PositiveInfinity; for (int i = 3; i < 40; i++) { double v = System.Math.Pow(0.1, i); A = Gaussian.FromMeanAndVariance(aMean, v); result2 = isProper ? GaussianProductOp.AAverageConditional(Product, A, B) : result; double diff = result.MaxDiff(result2); Console.WriteLine("{0}: {1} diff={2}", A, result2, diff.ToString("g4")); //Assert.True(diff <= prevDiff || diff < 1e-6); result2 = isProper ? GaussianProductOp_Slow.AAverageConditional(Product, A, B) : result; diff = result.MaxDiff(result2); Console.WriteLine("{0}: {1} diff={2}", A, result2, diff.ToString("g4")); Assert.True(diff <= prevDiff || diff < 1e-6); prevDiff = diff; } }
public static Gaussian DAverageConditional([SkipIfUniform] GammaPower exp, [Proper] Gaussian d) { double scale = 1 / exp.Power; Gaussian forward = GaussianProductOp.ProductAverageConditional(scale, d); Gaussian message = DAverageConditional(Gamma.FromNatural(exp.Shape - exp.Power, exp.Rate), forward); Gaussian backward = GaussianProductOp.BAverageConditional(message, scale); return(backward); }
/// <summary>Computations that depend on the observed value of numberOfIterationsDecreased and WeightPriors and FeatureIndexes and InstanceCount and InstanceFeatureCounts and FeatureValues and Labels and WeightConstraints</summary> /// <param name="numberOfIterations">The number of times to iterate each loop</param> private void Changed_numberOfIterationsDecreased_WeightPriors_FeatureIndexes_InstanceCount_InstanceFeatureCounts_7(int numberOfIterations) { if (this.Changed_numberOfIterationsDecreased_WeightPriors_FeatureIndexes_InstanceCount_InstanceFeatureCounts_7_iterationsDone == numberOfIterations) { return; } for (int iteration = this.Changed_numberOfIterationsDecreased_WeightPriors_FeatureIndexes_InstanceCount_InstanceFeatureCounts_7_iterationsDone; iteration < numberOfIterations; iteration++) { // Message to 'Weights_uses' from Replicate factor this.Weights_uses_B_toDef = ReplicateOp_Divide.ToDef <DistributionStructArray <Gaussian, double> >(this.Weights_uses_B, this.Weights_uses_B_toDef); // Message to 'Weights_uses' from Replicate factor this.Weights_uses_F_marginal = ReplicateOp_Divide.Marginal <DistributionStructArray <Gaussian, double> >(this.Weights_uses_B_toDef, this.weightPriors, this.Weights_uses_F_marginal); // Message to 'Weights_uses' from Replicate factor this.Weights_uses_F[1] = ReplicateOp_Divide.UsesAverageConditional <DistributionStructArray <Gaussian, double> >(this.Weights_uses_B[1], this.Weights_uses_F_marginal, 1, this.Weights_uses_F[1]); // Message to 'Weights_FeatureIndexes' from JaggedSubarray factor this.Weights_uses_F_1__marginal = JaggedSubarrayOp <double> .Marginal <DistributionStructArray <Gaussian, double>, Gaussian, object, DistributionStructArray <Gaussian, double> >(this.Weights_uses_F[1], this.IndexedWeights_B, this.featureIndexes, this.Weights_uses_F_1__marginal); for (int InstanceRange = 0; InstanceRange < this.instanceCount; InstanceRange++) { // Message to 'Weights_FeatureIndexes' from JaggedSubarray factor this.Weights_FeatureIndexes_F[InstanceRange] = JaggedSubarrayOp <double> .ItemsAverageConditional <DistributionStructArray <Gaussian, double>, Gaussian, DistributionStructArray <Gaussian, double> >(this.IndexedWeights_B[InstanceRange], this.Weights_uses_F[1], this.Weights_uses_F_1__marginal, this.featureIndexes, InstanceRange, this.Weights_FeatureIndexes_F[InstanceRange]); for (int InstanceFeatureRanges = 0; InstanceFeatureRanges < this.instanceFeatureCounts[InstanceRange]; InstanceFeatureRanges++) { // Message to 'FeatureScores' from Product factor this.FeatureScores_F[InstanceRange][InstanceFeatureRanges] = GaussianProductOp.ProductAverageConditional(this.featureValues[InstanceRange][InstanceFeatureRanges], this.Weights_FeatureIndexes_F[InstanceRange][InstanceFeatureRanges]); } // Message to 'Score' from Sum factor this.Score_F[InstanceRange] = FastSumOp.SumAverageConditional(this.FeatureScores_F[InstanceRange]); // Message to 'NoisyScore' from GaussianFromMeanAndVariance factor this.NoisyScore_F[InstanceRange] = GaussianFromMeanAndVarianceOp.SampleAverageConditional(this.Score_F[InstanceRange], 1.0); // Message to 'NoisyScore_use' from IsPositive factor this.NoisyScore_use_B[InstanceRange] = IsPositiveOp_Proper.XAverageConditional(Bernoulli.PointMass(this.labels[InstanceRange]), this.NoisyScore_F[InstanceRange]); // Message to 'Score' from GaussianFromMeanAndVariance factor this.Score_B[InstanceRange] = GaussianFromMeanAndVarianceOp.MeanAverageConditional(this.NoisyScore_use_B[InstanceRange], 1.0); // Message to 'FeatureScores' from Sum factor this.FeatureScores_B[InstanceRange] = FastSumOp.ArrayAverageConditional <DistributionStructArray <Gaussian, double> >(this.Score_B[InstanceRange], this.Score_F[InstanceRange], this.FeatureScores_F[InstanceRange], this.FeatureScores_B[InstanceRange]); for (int InstanceFeatureRanges = 0; InstanceFeatureRanges < this.instanceFeatureCounts[InstanceRange]; InstanceFeatureRanges++) { // Message to 'IndexedWeights' from Product factor this.IndexedWeights_B[InstanceRange][InstanceFeatureRanges] = GaussianProductOp.BAverageConditional(this.FeatureScores_B[InstanceRange][InstanceFeatureRanges], this.featureValues[InstanceRange][InstanceFeatureRanges]); } this.Weights_uses_F_1__marginal = JaggedSubarrayOp <double> .MarginalIncrement <DistributionStructArray <Gaussian, double>, Gaussian, DistributionStructArray <Gaussian, double> >(this.Weights_uses_F_1__marginal, this.Weights_FeatureIndexes_F[InstanceRange], this.IndexedWeights_B[InstanceRange], this.featureIndexes, InstanceRange); } // Message to 'Weights_uses' from JaggedSubarray factor this.Weights_uses_B[1] = JaggedSubarrayOp <double> .ArrayAverageConditional <Gaussian, DistributionStructArray <Gaussian, double>, DistributionStructArray <Gaussian, double> >(this.IndexedWeights_B, this.featureIndexes, this.Weights_uses_B[1]); this.OnProgressChanged(new ProgressChangedEventArgs(iteration)); } // Message to 'Weights_uses' from Replicate factor this.Weights_uses_B_toDef = ReplicateOp_Divide.ToDef <DistributionStructArray <Gaussian, double> >(this.Weights_uses_B, this.Weights_uses_B_toDef); // Message to 'Weights_marginal' from Variable factor this.Weights_marginal_F = VariableOp.MarginalAverageConditional <DistributionStructArray <Gaussian, double> >(this.Weights_uses_B_toDef, this.weightPriors, this.Weights_marginal_F); this.Changed_numberOfIterationsDecreased_WeightPriors_FeatureIndexes_InstanceCount_InstanceFeatureCounts_7_iterationsDone = numberOfIterations; }
/// <summary>Computations that depend on the observed value of xValueCount and xValues and wPrior and xIndices</summary> public void Changed_xValueCount_xValues_wPrior_xIndices() { if (this.Changed_xValueCount_xValues_wPrior_xIndices_iterationsDone == 1) { return; } for (int userFeature = 0; userFeature < this.XValueCount; userFeature++) { this.product_F[userFeature] = GaussianProductOp.ProductAverageConditional(this.XValues[userFeature], this.wSparse_F[userFeature]); } this.score_F = FastSumOp.SumAverageConditional(this.product_F); this.Changed_xValueCount_xValues_wPrior_xIndices_iterationsDone = 1; this.Changed_biasPrior_xValueCount_xValues_wPrior_xIndices_iterationsDone = 0; this.Changed_y_biasPrior_xValueCount_xValues_wPrior_xIndices_iterationsDone = 0; }
/// <summary> /// Computes the inner product of the feature weights and values. /// </summary> /// <param name="weights">The feature weights.</param> /// <param name="nonZeroValues">The sparse feature values.</param> /// <param name="nonZeroIndices">The sparse feature indices.</param> /// <returns>The contribution of the features.</returns> private static Gaussian ComputeFeatureContribution(IList <Gaussian> weights, IList <double> nonZeroValues, IList <int> nonZeroIndices) { Debug.Assert(nonZeroValues.Count == nonZeroIndices.Count, "The number of values must be equal to the number of indices."); int count = nonZeroValues.Count; var nonZeroWeights = new Gaussian[count]; SubarrayOp <double> .ItemsAverageConditional(weights, nonZeroIndices, nonZeroWeights); var products = new List <Gaussian>(count); for (int i = 0; i < count; ++i) { products.Add(GaussianProductOp.ProductAverageConditional(nonZeroWeights[i], nonZeroValues[i])); } return(FastSumOp.SumAverageConditional(products)); }
public void GaussianProductOp_ProductPointMassTest() { Gaussian A = new Gaussian(1, 2); Gaussian B = new Gaussian(3, 4); Gaussian pointMass = Gaussian.PointMass(4); Gaussian to_pointMass = GaussianProductOp.ProductAverageConditional(pointMass, A, B); double prevDiff = double.PositiveInfinity; for (int i = 0; i < 100; i++) { Gaussian Product = Gaussian.FromMeanAndVariance(4, System.Math.Pow(10, -i)); Gaussian to_product = GaussianProductOp.ProductAverageConditional(Product, A, B); double evidence = GaussianProductOp.LogEvidenceRatio(Product, A, B, to_product); Console.WriteLine($"{Product} {to_product} {evidence}"); double diff = to_product.MaxDiff(to_pointMass); Assert.True(diff <= prevDiff || diff < 1e-6); prevDiff = diff; } }
/// <summary>Computations that depend on the observed value of y and biasPrior and xValueCount and xValues and wPrior and xIndices</summary> public void Changed_y_biasPrior_xValueCount_xValues_wPrior_xIndices() { if (this.Changed_y_biasPrior_xValueCount_xValues_wPrior_xIndices_iterationsDone == 1) { return; } this.vdouble13_use_B = IsPositiveOp.XAverageConditional(this.Y, this.vdouble13_F); this.vdouble11_B = GaussianFromMeanAndVarianceOp.MeanAverageConditional(this.vdouble13_use_B, 1); this.bias_use_B = DoublePlusOp.BAverageConditional(this.vdouble11_B, this.score_F); this.bias_marginal_F = VariableOp.MarginalAverageConditional <Gaussian>(this.bias_use_B, this.BiasPrior, this.bias_marginal_F); this.score_B = DoublePlusOp.AAverageConditional(this.vdouble11_B, this.BiasPrior); this.product_B = FastSumOp.ArrayAverageConditional <DistributionStructArray <Gaussian, double> >(this.score_B, this.score_F, this.product_F, this.product_B); for (int userFeature = 0; userFeature < this.XValueCount; userFeature++) { this.wSparse_use_B[userFeature] = GaussianProductOp.BAverageConditional(this.product_B[userFeature], this.XValues[userFeature]); } this.wSparse_marginal_F = DerivedVariableOp.MarginalAverageConditional <DistributionStructArray <Gaussian, double> >(this.wSparse_use_B, this.wSparse_F, this.wSparse_marginal_F); this.Changed_y_biasPrior_xValueCount_xValues_wPrior_xIndices_iterationsDone = 1; }
/// <summary>Computations that depend on the observed value of InstanceCount and FeatureCount and FeatureValues and numberOfIterationsDecreased and WeightPriors and WeightConstraints</summary> private void Changed_InstanceCount_FeatureCount_FeatureValues_numberOfIterationsDecreased_WeightPriors_WeightCons10() { if (this.Changed_InstanceCount_FeatureCount_FeatureValues_numberOfIterationsDecreased_WeightPriors_WeightCons10_iterationsDone == 1) { return; } for (int InstanceRange = 0; InstanceRange < this.instanceCount; InstanceRange++) { for (int FeatureRange = 0; FeatureRange < this.featureCount; FeatureRange++) { // Message to 'FeatureScores' from Product factor this.FeatureScores_F[InstanceRange][FeatureRange] = GaussianProductOp.ProductAverageConditional(this.featureValues[InstanceRange][FeatureRange], this.Weights_depth1_rep_F[FeatureRange][InstanceRange]); } // Message to 'Score' from Sum factor this.Score_F[InstanceRange] = FastSumOp.SumAverageConditional(this.FeatureScores_F[InstanceRange]); // Message to 'NoisyScore' from GaussianFromMeanAndVariance factor this.NoisyScore_F[InstanceRange] = GaussianFromMeanAndVarianceOp.SampleAverageConditional(this.Score_F[InstanceRange], 1.0); // Message to 'Labels' from IsPositive factor this.Labels_F[InstanceRange] = IsPositiveOp.IsPositiveAverageConditional(this.NoisyScore_F[InstanceRange]); // Message to 'Labels_marginal' from DerivedVariable factor this.Labels_marginal_F[InstanceRange] = DerivedVariableOp.MarginalAverageConditional <Bernoulli>(this.Labels_use_B[InstanceRange], this.Labels_F[InstanceRange], this.Labels_marginal_F[InstanceRange]); } this.Changed_InstanceCount_FeatureCount_FeatureValues_numberOfIterationsDecreased_WeightPriors_WeightCons10_iterationsDone = 1; }
/// <include file='FactorDocs.xml' path='factor_docs/message_op_class[@name="RatioGaussianOp"]/message_doc[@name="LogAverageFactor(double, Gaussian, double)"]/*'/> public static double LogAverageFactor(double ratio, Gaussian a, double b) { Gaussian to_ratio = GaussianProductOp.AAverageConditional(a, b); return(to_ratio.GetLogProb(ratio)); }
/// <include file='FactorDocs.xml' path='factor_docs/message_op_class[@name="RatioGaussianOp"]/message_doc[@name="AAverageConditional(Gaussian, double)"]/*'/> public static Gaussian AAverageConditional([SkipIfUniform] Gaussian ratio, double b) { return(GaussianProductOp.ProductAverageConditional(ratio, b)); }
/// <include file='FactorDocs.xml' path='factor_docs/message_op_class[@name="RatioGaussianVmpOp"]/message_doc[@name="AAverageLogarithm(double, double)"]/*'/> public static Gaussian AAverageLogarithm(double ratio, double B) { return(GaussianProductOp.ProductAverageConditional(ratio, B)); }
/// <include file='FactorDocs.xml' path='factor_docs/message_op_class[@name="RatioGaussianVmpOp"]/message_doc[@name="RatioAverageLogarithm(Gaussian, double)"]/*'/> public static Gaussian RatioAverageLogarithm([SkipIfUniform] Gaussian A, double B) { return(GaussianProductOp.AAverageConditional(A, B)); }
/// <summary>Computations that do not depend on observed values</summary> private void Constant() { if (this.Constant_isDone) { return; } Gaussian t1_F = default(Gaussian); this.t1_marginal_F = Gaussian.Uniform(); Gaussian t1_use_B = Gaussian.Uniform(); // Message to 't1' from GaussianFromMeanAndVariance factor t1_F = GaussianFromMeanAndVarianceOp.SampleAverageConditional(1.0, 1.0); Gaussian[] t1_uses_F; Gaussian[] t1_uses_B; // Create array for 't1_uses' Forwards messages. t1_uses_F = new Gaussian[2]; // Create array for 't1_uses' Backwards messages. t1_uses_B = new Gaussian[2]; t1_uses_B[1] = Gaussian.Uniform(); t1_uses_B[0] = Gaussian.Uniform(); t1_uses_F[1] = Gaussian.Uniform(); t1_uses_F[0] = Gaussian.Uniform(); // Message to 't1_marginal' from Variable factor this.t1_marginal_F = VariableOp.MarginalAverageConditional <Gaussian>(t1_use_B, t1_F, this.t1_marginal_F); // Message to 't1_uses' from Replicate factor t1_uses_F[0] = ReplicateOp_NoDivide.UsesAverageConditional <Gaussian>(t1_uses_B, t1_F, 0, t1_uses_F[0]); // Message to 't1_uses' from Replicate factor t1_uses_F[1] = ReplicateOp_NoDivide.UsesAverageConditional <Gaussian>(t1_uses_B, t1_F, 1, t1_uses_F[1]); Gaussian t2_F = default(Gaussian); this.t2_marginal_F = Gaussian.Uniform(); Gaussian t2_use_B = Gaussian.Uniform(); // Message to 't2' from GaussianFromMeanAndVariance factor t2_F = GaussianFromMeanAndVarianceOp.SampleAverageConditional(10.0, 1.0); Gaussian[] t2_uses_F; Gaussian[] t2_uses_B; // Create array for 't2_uses' Forwards messages. t2_uses_F = new Gaussian[2]; // Create array for 't2_uses' Backwards messages. t2_uses_B = new Gaussian[2]; t2_uses_B[1] = Gaussian.Uniform(); t2_uses_B[0] = Gaussian.Uniform(); t2_uses_F[1] = Gaussian.Uniform(); t2_uses_F[0] = Gaussian.Uniform(); // Message to 't2_marginal' from Variable factor this.t2_marginal_F = VariableOp.MarginalAverageConditional <Gaussian>(t2_use_B, t2_F, this.t2_marginal_F); // Message to 't2_uses' from Replicate factor t2_uses_F[0] = ReplicateOp_NoDivide.UsesAverageConditional <Gaussian>(t2_uses_B, t2_F, 0, t2_uses_F[0]); // Message to 't2_uses' from Replicate factor t2_uses_F[1] = ReplicateOp_NoDivide.UsesAverageConditional <Gaussian>(t2_uses_B, t2_F, 1, t2_uses_F[1]); Gaussian vdouble6_F = default(Gaussian); this.vdouble6_marginal_F = Gaussian.Uniform(); Gaussian vdouble6_use_B = Gaussian.Uniform(); // Message to 'vdouble6' from Plus factor vdouble6_F = DoublePlusOp.SumAverageConditional(t1_uses_F[0], t2_uses_F[0]); // Message to 'vdouble6_marginal' from DerivedVariable factor this.vdouble6_marginal_F = DerivedVariableOp.MarginalAverageConditional <Gaussian>(vdouble6_use_B, vdouble6_F, this.vdouble6_marginal_F); Gaussian t4_F = default(Gaussian); this.t4_marginal_F = Gaussian.Uniform(); Gaussian t4_use_B = Gaussian.Uniform(); // Message to 't4' from GaussianFromMeanAndVariance factor t4_F = GaussianFromMeanAndVarianceOp.SampleAverageConditional(2.0, 1.0); Gaussian[] t4_uses_F; Gaussian[] t4_uses_B; // Create array for 't4_uses' Forwards messages. t4_uses_F = new Gaussian[2]; // Create array for 't4_uses' Backwards messages. t4_uses_B = new Gaussian[2]; t4_uses_B[1] = Gaussian.Uniform(); t4_uses_B[0] = Gaussian.Uniform(); t4_uses_F[1] = Gaussian.Uniform(); t4_uses_F[0] = Gaussian.Uniform(); // Message to 't4_marginal' from Variable factor this.t4_marginal_F = VariableOp.MarginalAverageConditional <Gaussian>(t4_use_B, t4_F, this.t4_marginal_F); // Message to 't4_uses' from Replicate factor t4_uses_F[0] = ReplicateOp_NoDivide.UsesAverageConditional <Gaussian>(t4_uses_B, t4_F, 0, t4_uses_F[0]); // Message to 't4_uses' from Replicate factor t4_uses_F[1] = ReplicateOp_NoDivide.UsesAverageConditional <Gaussian>(t4_uses_B, t4_F, 1, t4_uses_F[1]); Gaussian vdouble10_F = default(Gaussian); this.vdouble10_marginal_F = Gaussian.Uniform(); Gaussian vdouble10_use_B = Gaussian.Uniform(); // Message to 'vdouble10' from Plus factor vdouble10_F = DoublePlusOp.SumAverageConditional(t1_uses_F[1], t2_uses_F[1]); // Message to 'vdouble10_marginal' from DerivedVariable factor this.vdouble10_marginal_F = DerivedVariableOp.MarginalAverageConditional <Gaussian>(vdouble10_use_B, vdouble10_F, this.vdouble10_marginal_F); Gaussian t5_F = default(Gaussian); this.t5_marginal_F = Gaussian.Uniform(); Gaussian t5_use_B = Gaussian.Uniform(); // Message to 't5' from Product factor t5_F = GaussianProductOp.ProductAverageConditional(t5_use_B, vdouble10_F, t4_uses_F[0]); // Message to 't5_marginal' from DerivedVariable factor this.t5_marginal_F = DerivedVariableOp.MarginalAverageConditional <Gaussian>(t5_use_B, t5_F, this.t5_marginal_F); Gaussian t7_F = default(Gaussian); this.t7_marginal_F = Gaussian.Uniform(); Gaussian t7_use_B = Gaussian.Uniform(); // Message to 't7' from Product factor t7_F = GaussianProductOp.ProductAverageConditional(t7_use_B, vdouble6_F, t4_uses_F[1]); // Message to 't7_marginal' from DerivedVariable factor this.t7_marginal_F = DerivedVariableOp.MarginalAverageConditional <Gaussian>(t7_use_B, t7_F, this.t7_marginal_F); this.Constant_isDone = true; }
public void ProductOpTest() { Assert.True(GaussianProductVmpOp.ProductAverageLogarithm( 2.0, Gaussian.FromMeanAndVariance(3.0, 5.0)).MaxDiff(Gaussian.FromMeanAndVariance(2 * 3, 4 * 5)) < 1e-8); Assert.True(GaussianProductOp.ProductAverageConditional( 2.0, Gaussian.FromMeanAndVariance(3.0, 5.0)).MaxDiff(Gaussian.FromMeanAndVariance(2 * 3, 4 * 5)) < 1e-8); Assert.True(GaussianProductOp.ProductAverageConditional(new Gaussian(0, 1), Gaussian.PointMass(2.0), Gaussian.FromMeanAndVariance(3.0, 5.0)).MaxDiff(Gaussian.FromMeanAndVariance(2 * 3, 4 * 5)) < 1e-8); Assert.True(GaussianProductVmpOp.ProductAverageLogarithm( 0.0, Gaussian.FromMeanAndVariance(3.0, 5.0)).MaxDiff(Gaussian.PointMass(0.0)) < 1e-8); Assert.True(GaussianProductOp.ProductAverageConditional( 0.0, Gaussian.FromMeanAndVariance(3.0, 5.0)).MaxDiff(Gaussian.PointMass(0.0)) < 1e-8); Assert.True(GaussianProductOp.ProductAverageConditional(new Gaussian(0, 1), Gaussian.PointMass(0.0), Gaussian.FromMeanAndVariance(3.0, 5.0)).MaxDiff(Gaussian.PointMass(0.0)) < 1e-8); Assert.True(GaussianProductVmpOp.ProductAverageLogarithm( Gaussian.FromMeanAndVariance(2, 4), Gaussian.FromMeanAndVariance(3, 5)).MaxDiff(Gaussian.FromMeanAndVariance(2 * 3, 4 * 5 + 3 * 3 * 4 + 2 * 2 * 5)) < 1e-8); Assert.True(GaussianProductOp.ProductAverageConditional(Gaussian.Uniform(), Gaussian.FromMeanAndVariance(2, 4), Gaussian.FromMeanAndVariance(3, 5)).MaxDiff(Gaussian.FromMeanAndVariance(2 * 3, 4 * 5 + 3 * 3 * 4 + 2 * 2 * 5)) < 1e-8); Assert.True(GaussianProductOp.ProductAverageConditional(Gaussian.FromMeanAndVariance(0, 1e16), Gaussian.FromMeanAndVariance(2, 4), Gaussian.FromMeanAndVariance(3, 5)).MaxDiff(Gaussian.FromMeanAndVariance(2 * 3, 4 * 5 + 3 * 3 * 4 + 2 * 2 * 5)) < 1e-4); Assert.True(GaussianProductOp.AAverageConditional(6.0, 2.0) .MaxDiff(Gaussian.PointMass(6.0 / 2.0)) < 1e-8); Assert.True(GaussianProductOp.AAverageConditional(6.0, new Gaussian(1, 3), Gaussian.PointMass(2.0)) .MaxDiff(Gaussian.PointMass(6.0 / 2.0)) < 1e-8); Assert.True(GaussianProductOp.AAverageConditional(0.0, 0.0).IsUniform()); Assert.True(GaussianProductOp.AAverageConditional(Gaussian.Uniform(), 2.0).IsUniform()); Assert.True(GaussianProductOp.AAverageConditional(Gaussian.Uniform(), new Gaussian(1, 3), Gaussian.PointMass(2.0)).IsUniform()); Assert.True(GaussianProductOp.AAverageConditional(Gaussian.Uniform(), new Gaussian(1, 3), new Gaussian(2, 4)).IsUniform()); Gaussian aPrior = Gaussian.FromMeanAndVariance(0.0, 1000.0); Assert.True((GaussianProductOp.AAverageConditional( Gaussian.FromMeanAndVariance(10.0, 1.0), aPrior, Gaussian.FromMeanAndVariance(5.0, 1.0)) * aPrior).MaxDiff( Gaussian.FromMeanAndVariance(2.208041421368822, 0.424566765678152)) < 1e-4); Gaussian g = new Gaussian(0, 1); Assert.True(GaussianProductOp.AAverageConditional(g, 0.0).IsUniform()); Assert.True(GaussianProductOp.AAverageConditional(0.0, 0.0).IsUniform()); Assert.True(GaussianProductVmpOp.AAverageLogarithm(g, 0.0).IsUniform()); Assert.True(Gaussian.PointMass(3.0).MaxDiff(GaussianProductVmpOp.AAverageLogarithm(6.0, 2.0)) < 1e-10); Assert.True(GaussianProductVmpOp.AAverageLogarithm(0.0, 0.0).IsUniform()); try { Assert.True(GaussianProductVmpOp.AAverageLogarithm(6.0, g).IsUniform()); Assert.True(false, "Did not throw NotSupportedException"); } catch (NotSupportedException) { } try { g = GaussianProductOp.AAverageConditional(12.0, 0.0); Assert.True(false, "Did not throw AllZeroException"); } catch (AllZeroException) { } try { g = GaussianProductVmpOp.AAverageLogarithm(12.0, 0.0); Assert.True(false, "Did not throw AllZeroException"); } catch (AllZeroException) { } }