public void SingleDependencyBeforeUpdateIsAsSmallAsPossible()
{
GC.Collect();
long start = GC.GetTotalMemory(true);
Independent <int> newIndependent = new Independent <int>();
Dependent <int> newDependent = new Dependent <int>(() => newIndependent);
newIndependent.Value = 42;
long end = GC.GetTotalMemory(true);
// Started at 336.
// Making Precedent a base class: 312.
// Removing Gain/LoseDependent events: 288.
// Making IsUpToDate no longer a precident: 248.
// Custom linked list implementation for dependents: 200.
// Custom linked list implementation for precedents: 160.
// Other optimizations: 144.
// Added WeakReferenceToSelf: 148.
// Removed WeakReferenceToSelf: 124.
Assert.AreEqual(124 + IndependentPlatformOffset, end - start);
int value = newDependent;
Assert.AreEqual(42, value);
}
public void ConstructorTest()
{
var p1 = new NormalDistribution(4.2, 1);
var p2 = new NormalDistribution(7.0, 2);
Independent <NormalDistribution> target = new Independent <NormalDistribution>(p1, p2);
Assert.AreEqual(target.Components[0], p1);
Assert.AreEqual(target.Components[1], p2);
Assert.AreEqual(2, target.Dimension);
Assert.AreEqual(4.2, target.Mean[0]);
Assert.AreEqual(7.0, target.Mean[1]);
Assert.AreEqual(1, target.Variance[0]);
Assert.AreEqual(4, target.Variance[1]);
Assert.AreEqual(1, target.Covariance[0, 0]);
Assert.AreEqual(4, target.Covariance[1, 1]);
Assert.AreEqual(0, target.Covariance[0, 1]);
Assert.AreEqual(0, target.Covariance[1, 0]);
var text = target.ToString("N2", System.Globalization.CultureInfo.InvariantCulture);
Assert.AreEqual("Independent(x0, x1; N(x0; μ = 4.20, σ² = 1.00) + N(x1; μ = 7.00, σ² = 4.00))", text);
}
public void ProbabilityFunctionTest()
{
var p1 = new NormalDistribution(4.2, 1);
var p2 = new NormalDistribution(7.0, 2);
Independent <NormalDistribution> target = new Independent <NormalDistribution>(p1, p2);
double[] x;
double actual, expected;
x = new double[] { 4.2, 7.0 };
actual = target.ProbabilityDensityFunction(x);
expected = p1.ProbabilityDensityFunction(x[0]) * p2.ProbabilityDensityFunction(x[1]);
Assert.AreEqual(expected, actual, 1e-10);
Assert.IsFalse(double.IsNaN(actual));
x = new double[] { 0.0, 0.0 };
actual = target.ProbabilityDensityFunction(x);
expected = p1.ProbabilityDensityFunction(x[0]) * p2.ProbabilityDensityFunction(x[1]);
Assert.AreEqual(expected, actual, 1e-10);
Assert.IsFalse(double.IsNaN(actual));
x = new double[] { 7.0, 4.2 };
actual = target.ProbabilityDensityFunction(x);
expected = p1.ProbabilityDensityFunction(x[0]) * p2.ProbabilityDensityFunction(x[1]);
Assert.AreEqual(expected, actual, 1e-10);
Assert.IsFalse(double.IsNaN(actual));
}
public void CumulativeFunctionTest()
{
var p1 = new NormalDistribution(4.2, 1);
var p2 = new NormalDistribution(7.0, 2);
Independent <NormalDistribution> target = new Independent <NormalDistribution>(p1, p2);
double[] x;
double actual, expected;
x = new double[] { 4.2, 7.0 };
actual = target.DistributionFunction(x);
expected = p1.DistributionFunction(x[0]) * p2.DistributionFunction(x[1]);
Assert.AreEqual(expected, actual);
x = new double[] { 0.0, 0.0 };
actual = target.DistributionFunction(x);
expected = p1.DistributionFunction(x[0]) * p2.DistributionFunction(x[1]);
Assert.AreEqual(expected, actual);
x = new double[] { 7.0, 4.2 };
actual = target.DistributionFunction(x);
expected = p1.DistributionFunction(x[0]) * p2.DistributionFunction(x[1]);
Assert.AreEqual(expected, actual);
}
public static HiddenMarkovClassifier <Independent> CreateModel3()
{
// Create a Continuous density Hidden Markov Model Sequence Classifier
// to detect a multivariate sequence and the same sequence backwards.
var comp1 = new GeneralDiscreteDistribution(3);
var comp2 = new NormalDistribution(1);
var comp3 = new NormalDistribution(2);
var comp4 = new NormalDistribution(3);
var comp5 = new NormalDistribution(4);
var density = new Independent(comp1, comp2, comp3, comp4, comp5);
// Creates a sequence classifier containing 2 hidden Markov Models with 2 states
// and an underlying multivariate mixture of Normal distributions as density.
var classifier = new HiddenMarkovClassifier <Independent>(
2, new Forward(5), density);
// Configure the learning algorithms to train the sequence classifier
var teacher = new HiddenMarkovClassifierLearning <Independent>(
classifier,
// Train each model until the log-likelihood changes less than 0.0001
modelIndex => new BaumWelchLearning <Independent>(
classifier.Models[modelIndex])
{
Tolerance = 0.0001,
Iterations = 0,
}
);
// Train the sequence classifier using the algorithm
double logLikelihood = teacher.Run(sequences2, labels2);
return(classifier);
}
public void ConstructorTest()
{
var p1 = new NormalDistribution(4.2, 1);
var p2 = new NormalDistribution(7.0, 2);
Independent<NormalDistribution> target = new Independent<NormalDistribution>(p1, p2);
Assert.AreEqual(target.Components[0], p1);
Assert.AreEqual(target.Components[1], p2);
Assert.AreEqual(2, target.Dimension);
Assert.AreEqual(4.2, target.Mean[0]);
Assert.AreEqual(7.0, target.Mean[1]);
Assert.AreEqual(1, target.Variance[0]);
Assert.AreEqual(4, target.Variance[1]);
Assert.AreEqual(1, target.Covariance[0, 0]);
Assert.AreEqual(4, target.Covariance[1, 1]);
Assert.AreEqual(0, target.Covariance[0, 1]);
Assert.AreEqual(0, target.Covariance[1, 0]);
var text = target.ToString("N2", System.Globalization.CultureInfo.InvariantCulture);
Assert.AreEqual("Independent(x0, x1; N(x0; μ = 4.20, σ² = 1.00) + N(x1; μ = 7.00, σ² = 4.00))", text);
}
public void FitTest()
{
double[][] data =
{
new double[] { 1, 8 },
new double[] { 2, 6 },
new double[] { 5, 7 },
new double[] { 3, 9 },
};
var p1 = new NormalDistribution(0, 1);
var p2 = new NormalDistribution(0, 1);
Independent <NormalDistribution> target = new Independent <NormalDistribution>(p1, p2);
target.Fit(data);
Assert.AreEqual(2.75, target.Mean[0]);
Assert.AreEqual(7.50, target.Mean[1]);
Assert.AreEqual(2.91666, target.Variance[0], 1e-5);
Assert.AreEqual(1.66666, target.Variance[1], 1e-5);
Assert.AreEqual(target.Variance[0], target.Covariance[0, 0]);
Assert.AreEqual(target.Variance[1], target.Covariance[1, 1]);
Assert.AreEqual(0, target.Covariance[0, 1]);
Assert.AreEqual(0, target.Covariance[1, 0]);
}
public void ProbabilityFunctionTest()
{
var p1 = new NormalDistribution(4.2, 1);
var p2 = new NormalDistribution(7.0, 2);
Independent<NormalDistribution> target = new Independent<NormalDistribution>(p1, p2);
double[] x;
double actual, expected;
x = new double[] { 4.2, 7.0 };
actual = target.ProbabilityDensityFunction(x);
expected = p1.ProbabilityDensityFunction(x[0]) * p2.ProbabilityDensityFunction(x[1]);
Assert.AreEqual(expected, actual, 1e-10);
Assert.IsFalse(double.IsNaN(actual));
x = new double[] { 0.0, 0.0 };
actual = target.ProbabilityDensityFunction(x);
expected = p1.ProbabilityDensityFunction(x[0]) * p2.ProbabilityDensityFunction(x[1]);
Assert.AreEqual(expected, actual, 1e-10);
Assert.IsFalse(double.IsNaN(actual));
x = new double[] { 7.0, 4.2 };
actual = target.ProbabilityDensityFunction(x);
expected = p1.ProbabilityDensityFunction(x[0]) * p2.ProbabilityDensityFunction(x[1]);
Assert.AreEqual(expected, actual, 1e-10);
Assert.IsFalse(double.IsNaN(actual));
}
public void NumberOfFeaturesTest()
{
Independent initial = new Independent(
new GeneralDiscreteDistribution(46), // output,
new NormalDistribution(0, 1), // headAngle,
new NormalDistribution(0, 1), // handAngle,
new NormalDistribution(0, 1), // relHandX,
new NormalDistribution(0, 1) // relHandY,
);
var model = new HiddenMarkovClassifier <Independent>(
classes: 13, topology: new Forward(5), initial: initial);
var function = new MarkovMultivariateFunction(model);
var field = new HiddenConditionalRandomField <double[]>(function);
int discreteDistributions = 1;
int continuousDistributions = 4;
int symbols = 46;
int states = 5;
int classes = 13;
int expected = classes *
(1 + states + states * states +
states * discreteDistributions * symbols +
states * continuousDistributions * 3);
Assert.AreEqual(expected, field.Function.Features.Length);
Assert.AreEqual(expected, field.Function.Weights.Length);
Assert.AreEqual(4173, expected);
}
public void IndependentModelExample()
{
double[][] samples =
{
new double[] { 0, 1 },
new double[] { 1, 2 },
new double[] { 1, 1 },
new double[] { 0, 7 },
new double[] { 1, 1 },
new double[] { 6, 2 },
new double[] { 6, 5 },
new double[] { 5, 1 },
new double[] { 7, 1 },
new double[] { 5, 1 }
};
// Create a new 2D independent Normal distribution
var independent = new Independent <NormalDistribution>(
new NormalDistribution(), new NormalDistribution());
// Estimate it!
independent.Fit(samples);
var a = independent.Components[0].ToString("N2"); // N(x1; μ = 3.20, σ² = 7.96)
var b = independent.Components[1].ToString("N2"); // N(x2; μ = 2.20, σ² = 4.40)
Assert.AreEqual("N(x; μ = 3.20, σ² = 7.96)", a);
Assert.AreEqual("N(x; μ = 2.20, σ² = 4.40)", b);
}
private const double dist = 0.0005396; //constant that is approximatively 60m in Earth coordinate
/// <summary>
/// Initializes a new instance of ClientCoordsGenerator class.
/// </summary>
/// <param name="nominatimClient">Nominatim client.</param>
/// <param name="standardDeviation">Standard deviation for two dimensional uniform distribution.</param>
public ClientCoordsGenerator(INominatimClient nominatimClient, double standardDeviation)
{
_nominatimClient = nominatimClient;
var uniform = new UniformContinuousDistribution(-dist, +dist);
_distribution = new Independent <UniformContinuousDistribution>(uniform, uniform);
}
public static HiddenMarkovClassifier <Independent> CreateModel2(out double[][][] sequences, out int[] labels)
{
sequences = new double[][][]
{
new double[][]
{
// This is the first sequence with label = 0
new double[] { 0, 1.1 },
new double[] { 1, 2.5 },
new double[] { 1, 3.4 },
new double[] { 1, 4.7 },
new double[] { 2, 5.8 },
},
new double[][]
{
// This is the second sequence with label = 1
new double[] { 2, 3.2 },
new double[] { 2, 2.6 },
new double[] { 1, 1.2 },
new double[] { 1, 0.8 },
new double[] { 0, 1.1 },
}
};
labels = new[] { 0, 1 };
// Create a Continuous density Hidden Markov Model Sequence Classifier
// to detect a multivariate sequence and the same sequence backwards.
var comp1 = new GeneralDiscreteDistribution(3);
var comp2 = new NormalDistribution(1);
var density = new Independent(comp1, comp2);
// Creates a sequence classifier containing 2 hidden Markov Models with 2 states
// and an underlying multivariate mixture of Normal distributions as density.
var classifier = new HiddenMarkovClassifier <Independent>(
2, new Ergodic(2), density);
// Configure the learning algorithms to train the sequence classifier
var teacher = new HiddenMarkovClassifierLearning <Independent>(
classifier,
// Train each model until the log-likelihood changes less than 0.0001
modelIndex => new BaumWelchLearning <Independent>(
classifier.Models[modelIndex])
{
Tolerance = 0.0001,
Iterations = 0,
}
);
// Train the sequence classifier using the algorithm
double logLikelihood = teacher.Run(sequences, labels);
Assert.AreEqual(double.NegativeInfinity, logLikelihood); // only one training sample per class
return(classifier);
}
public void FittingOptionsTest()
{
double[][] data1 =
{
new double[] { 0, 8 },
new double[] { 0, 6 },
new double[] { 0, 7 },
new double[] { 0, 9 },
};
double[][] data2 =
{
new double[] { 8, 0 },
new double[] { 6, 0 },
new double[] { 7, 0 },
new double[] { 9, 0 },
};
var p1 = new NormalDistribution(0, 1);
var p2 = new NormalDistribution(0, 1);
{
Independent <NormalDistribution> target = new Independent <NormalDistribution>(p1, p2);
target.Fit(data1, new IndependentOptions()
{
InnerOption = new NormalOptions()
{
Regularization = 1e-5
}
});
Assert.AreEqual(0.00, target.Mean[0]);
Assert.AreEqual(7.50, target.Mean[1]);
Assert.AreEqual(1e-5, target.Variance[0]);
Assert.AreEqual(1.66666, target.Variance[1], 1e-5);
}
{
Independent <NormalDistribution> target = new Independent <NormalDistribution>(p1, p2);
target.Fit(data2, new IndependentOptions()
{
InnerOption = new NormalOptions()
{
Regularization = 1e-5
}
});
Assert.AreEqual(7.5, target.Mean[0]);
Assert.AreEqual(0, target.Mean[1]);
Assert.AreEqual(1.66666, target.Variance[0], 1e-5);
Assert.AreEqual(1e-5, target.Variance[1]);
}
}
/// <summary>
/// Object hash code
/// </summary>
/// <returns></returns>
public override int GetHashCode()
{
unchecked
{
var hashCode = Independent.GetHashCode();
hashCode = (hashCode * 397) ^ (Enzymes != null ? Enzymes.GetHashCode() : 0);
return(hashCode);
}
}
public static HiddenMarkovClassifier<Independent> CreateModel1()
{
// Create a Continuous density Hidden Markov Model Sequence Classifier
// to detect a multivariate sequence and the same sequence backwards.
double[][][] sequences = new double[][][]
{
new double[][]
{
// This is the first sequence with label = 0
new double[] { 0 },
new double[] { 1 },
new double[] { 2 },
new double[] { 3 },
new double[] { 4 },
},
new double[][]
{
// This is the second sequence with label = 1
new double[] { 4 },
new double[] { 3 },
new double[] { 2 },
new double[] { 1 },
new double[] { 0 },
}
};
// Labels for the sequences
int[] labels = { 0, 1 };
// Creates a sequence classifier containing 2 hidden Markov Models
// with 2 states and an underlying Normal distribution as density.
NormalDistribution component = new NormalDistribution();
Independent density = new Independent(component);
var classifier = new HiddenMarkovClassifier<Independent>(2, new Ergodic(2), density);
// Configure the learning algorithms to train the sequence classifier
var teacher = new HiddenMarkovClassifierLearning<Independent>(classifier,
// Train each model until the log-likelihood changes less than 0.001
modelIndex => new BaumWelchLearning<Independent>(classifier.Models[modelIndex])
{
Tolerance = 0.0001,
Iterations = 0
}
);
// Train the sequence classifier using the algorithm
double logLikelihood = teacher.Run(sequences, labels);
Assert.AreEqual(-13.271981026832929d, logLikelihood);
return classifier;
}
public static HiddenMarkovClassifier <Independent> CreateModel1()
{
// Create a Continuous density Hidden Markov Model Sequence Classifier
// to detect a multivariate sequence and the same sequence backwards.
double[][][] sequences = new double[][][]
{
new double[][]
{
// This is the first sequence with label = 0
new double[] { 0 },
new double[] { 1 },
new double[] { 2 },
new double[] { 3 },
new double[] { 4 },
},
new double[][]
{
// This is the second sequence with label = 1
new double[] { 4 },
new double[] { 3 },
new double[] { 2 },
new double[] { 1 },
new double[] { 0 },
}
};
// Labels for the sequences
int[] labels = { 0, 1 };
// Creates a sequence classifier containing 2 hidden Markov Models
// with 2 states and an underlying Normal distribution as density.
NormalDistribution component = new NormalDistribution();
Independent density = new Independent(component);
var classifier = new HiddenMarkovClassifier <Independent>(2, new Ergodic(2), density);
// Configure the learning algorithms to train the sequence classifier
var teacher = new HiddenMarkovClassifierLearning <Independent>(classifier,
// Train each model until the log-likelihood changes less than 0.001
modelIndex => new BaumWelchLearning <Independent>(classifier.Models[modelIndex])
{
Tolerance = 0.0001,
Iterations = 0
}
);
// Train the sequence classifier using the algorithm
double logLikelihood = teacher.Run(sequences, labels);
Assert.AreEqual(-13.271981026832929d, logLikelihood, 1e-10);
return(classifier);
}
public void FittingNoOptionsTest()
{
double[][] data1 =
{
new double[] { 0, 8 },
new double[] { 0, 6 },
new double[] { 0, 7 },
new double[] { 0, 9 },
};
double[][] data2 =
{
new double[] { 8, 0 },
new double[] { 6, 0 },
new double[] { 7, 0 },
new double[] { 9, 0 },
};
var p1 = new NormalDistribution(0, 1);
var p2 = new NormalDistribution(0, 1);
bool thrown1 = false;
bool thrown2 = false;
try
{
var target = new Independent <NormalDistribution>(p1, p2);
target.Fit(data1);
}
catch (ArgumentException)
{
thrown1 = true;
}
try
{
var target = new Independent <NormalDistribution>(p1, p2);
target.Fit(data2);
}
catch (ArgumentException)
{
thrown2 = true;
}
Assert.IsTrue(thrown1);
Assert.IsTrue(thrown2);
}
public void IndependentIsAsSmallAsPossible()
{
GC.Collect();
long start = GC.GetTotalMemory(true);
Independent <int> newIndependent = new Independent <int>();
newIndependent.Value = 42;
long end = GC.GetTotalMemory(true);
// Started at 92.
// Making Precedent a base class: 80.
// Removing Gain/LoseDependent events: 72.
// Custom linked list implementation for dependents: 48.
// Other optimizations: 40.
// Removed WeakReferenceToSelf: 20.
Assert.AreEqual(20 + IndependentPlatformOffset, end - start);
int value = newIndependent;
Assert.AreEqual(42, value);
}
public void ConstructorTest()
{
var p1 = new NormalDistribution(4.2, 1);
var p2 = new NormalDistribution(7.0, 2);
Independent <NormalDistribution> target = new Independent <NormalDistribution>(p1, p2);
Assert.AreEqual(target.Components[0], p1);
Assert.AreEqual(target.Components[1], p2);
Assert.AreEqual(2, target.Dimension);
Assert.AreEqual(4.2, target.Mean[0]);
Assert.AreEqual(7.0, target.Mean[1]);
Assert.AreEqual(1, target.Variance[0]);
Assert.AreEqual(4, target.Variance[1]);
Assert.AreEqual(1, target.Covariance[0, 0]);
Assert.AreEqual(4, target.Covariance[1, 1]);
Assert.AreEqual(0, target.Covariance[0, 1]);
Assert.AreEqual(0, target.Covariance[1, 0]);
}
public void ConstructorTest()
{
var p1 = new NormalDistribution(4.2, 1);
var p2 = new NormalDistribution(7.0, 2);
Independent<NormalDistribution> target = new Independent<NormalDistribution>(p1, p2);
Assert.AreEqual(target.Components[0], p1);
Assert.AreEqual(target.Components[1], p2);
Assert.AreEqual(2, target.Dimension);
Assert.AreEqual(4.2, target.Mean[0]);
Assert.AreEqual(7.0, target.Mean[1]);
Assert.AreEqual(1, target.Variance[0]);
Assert.AreEqual(4, target.Variance[1]);
Assert.AreEqual(1, target.Covariance[0, 0]);
Assert.AreEqual(4, target.Covariance[1, 1]);
Assert.AreEqual(0, target.Covariance[0, 1]);
Assert.AreEqual(0, target.Covariance[1, 0]);
}
public void FittingOptionsTest()
{
double[][] data1 =
{
new double[] { 0, 8 },
new double[] { 0, 6 },
new double[] { 0, 7 },
new double[] { 0, 9 },
};
double[][] data2 =
{
new double[] { 8, 0 },
new double[] { 6, 0 },
new double[] { 7, 0 },
new double[] { 9, 0 },
};
var p1 = new NormalDistribution(0, 1);
var p2 = new NormalDistribution(0, 1);
{
Independent<NormalDistribution> target = new Independent<NormalDistribution>(p1, p2);
target.Fit(data1, new IndependentOptions()
{
InnerOption = new NormalOptions()
{
Regularization = 1e-5
}
});
Assert.AreEqual(0.00, target.Mean[0]);
Assert.AreEqual(7.50, target.Mean[1]);
Assert.AreEqual(1e-5, target.Variance[0]);
Assert.AreEqual(1.66666, target.Variance[1], 1e-5);
}
{
Independent<NormalDistribution> target = new Independent<NormalDistribution>(p1, p2);
target.Fit(data2, new IndependentOptions()
{
InnerOption = new NormalOptions()
{
Regularization = 1e-5
}
});
Assert.AreEqual(7.5, target.Mean[0]);
Assert.AreEqual(0, target.Mean[1]);
Assert.AreEqual(1.66666, target.Variance[0], 1e-5);
Assert.AreEqual(1e-5, target.Variance[1]);
}
}
public static HiddenMarkovClassifier <Independent <NormalDistribution> > CreateModel4(out double[][][] words, out int[] labels, bool usePriors)
{
double[][] hello =
{
new double[] { 1.0, 0.1, 0.0, 0.0 }, // let's say the word
new double[] { 0.0, 1.0, 0.1, 0.1 }, // hello took 6 frames
new double[] { 0.0, 1.0, 0.1, 0.1 }, // to be recorded.
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 0.1, 1.1 },
};
double[][] car =
{
new double[] { 0.0, 0.0, 0.0, 1.0 }, // the car word
new double[] { 0.1, 0.0, 1.0, 0.1 }, // took only 4.
new double[] { 0.0, 0.0, 0.1, 0.0 },
new double[] { 1.0, 0.0, 0.0, 0.0 },
};
double[][] wardrobe =
{
new double[] { 0.0, 0.0, 1.0, 0.0 }, // same for the
new double[] { 0.1, 0.0, 1.0, 0.1 }, // wardrobe word.
new double[] { 0.0, 0.1, 1.0, 0.0 },
new double[] { 0.1, 0.0, 1.0, 0.1 },
};
double[][] wardrobe2 =
{
new double[] { 0.0, 0.0, 1.0, 0.0 }, // same for the
new double[] { 0.2, 0.0, 1.0, 0.1 }, // wardrobe word.
new double[] { 0.0, 0.1, 1.0, 0.0 },
new double[] { 0.1, 0.0, 1.0, 0.2 },
};
words = new double[][][] { hello, car, wardrobe, wardrobe2 };
labels = new [] { 0, 1, 2, 2 };
var initial = new Independent <NormalDistribution>
(
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1)
);
int numberOfWords = 3;
int numberOfStates = 5;
var classifier = new HiddenMarkovClassifier <Independent <NormalDistribution> >
(
classes: numberOfWords,
topology: new Forward(numberOfStates),
initial: initial
);
var teacher = new HiddenMarkovClassifierLearning <Independent <NormalDistribution> >(classifier,
modelIndex => new BaumWelchLearning <Independent <NormalDistribution> >(classifier.Models[modelIndex])
{
Tolerance = 0.001,
Iterations = 100,
FittingOptions = new IndependentOptions()
{
InnerOption = new NormalOptions()
{
Regularization = 1e-5
}
}
}
);
if (usePriors)
{
teacher.Empirical = true;
}
double logLikelihood = teacher.Run(words, labels);
Assert.AreEqual(208.38345600145777d, logLikelihood);
return(classifier);
}
public static HiddenMarkovClassifier <Independent> CreateModel3(out double[][][] sequences2, out int[] labels2)
{
sequences2 = new double[][][]
{
new double[][]
{
// This is the first sequence with label = 0
new double[] { 1, 1.12, 2.41, 1.17, 9.3 },
new double[] { 1, 2.54, 1.45, 0.16, 4.5 },
new double[] { 1, 3.46, 2.63, 1.15, 9.2 },
new double[] { 1, 4.73, 0.41, 1.54, 5.5 },
new double[] { 2, 5.81, 2.42, 1.13, 9.1 },
},
new double[][]
{
// This is the first sequence with label = 0
new double[] { 0, 1.49, 2.48, 1.18, 9.37 },
new double[] { 1, 2.18, 1.44, 2.19, 1.56 },
new double[] { 1, 3.77, 2.62, 1.10, 9.25 },
new double[] { 2, 4.76, 5.44, 3.58, 5.54 },
new double[] { 2, 5.85, 2.46, 1.16, 5.13 },
new double[] { 2, 4.84, 5.44, 3.54, 5.52 },
new double[] { 2, 5.83, 3.41, 1.22, 5.11 },
},
new double[][]
{
// This is the first sequence with label = 0
new double[] { 2, 1.11, 2.41, 1.12, 2.31 },
new double[] { 1, 2.52, 3.73, 0.12, 4.50 },
new double[] { 1, 3.43, 2.61, 1.24, 9.29 },
new double[] { 1, 4.74, 2.42, 2.55, 6.57 },
new double[] { 2, 5.85, 2.43, 1.16, 9.16 },
},
new double[][]
{
// This is the second sequence with label = 1
new double[] { 0, 1.26, 5.44, 1.56, 9.55 },
new double[] { 2, 2.67, 5.45, 4.27, 1.54 },
new double[] { 1, 1.28, 3.46, 2.18, 4.13 },
new double[] { 1, 5.89, 2.57, 1.79, 5.02 },
new double[] { 0, 1.40, 2.48, 2.10, 6.41 },
},
new double[][]
{
// This is the second sequence with label = 1
new double[] { 2, 3.21, 2.49, 1.54, 9.17 },
new double[] { 2, 2.62, 5.40, 4.25, 1.54 },
new double[] { 1, 1.53, 6.49, 2.17, 4.52 },
new double[] { 1, 2.84, 2.58, 1.73, 6.04 },
new double[] { 1, 1.45, 2.47, 2.28, 5.42 },
new double[] { 1, 1.46, 2.46, 2.35, 5.41 },
},
new double[][]
{
// This is the second sequence with label = 1
new double[] { 1, 5.27, 5.45, 1.4, 9.5 },
new double[] { 2, 2.68, 2.54, 3.2, 2.2 },
new double[] { 1, 2.89, 3.83, 2.6, 4.1 },
new double[] { 1, 1.80, 1.32, 1.2, 4.2 },
new double[] { 0, 1.41, 2.41, 2.1, 6.4 },
}
};
labels2 = new[] { 0, 0, 0, 1, 1, 1 };
// Create a Continuous density Hidden Markov Model Sequence Classifier
// to detect a multivariate sequence and the same sequence backwards.
var comp1 = new GeneralDiscreteDistribution(3);
var comp2 = new NormalDistribution(1);
var comp3 = new NormalDistribution(2);
var comp4 = new NormalDistribution(3);
var comp5 = new NormalDistribution(4);
var density = new Independent(comp1, comp2, comp3, comp4, comp5);
// Creates a sequence classifier containing 2 hidden Markov Models with 2 states
// and an underlying multivariate mixture of Normal distributions as density.
var classifier = new HiddenMarkovClassifier <Independent>(
2, new Forward(5), density);
// Configure the learning algorithms to train the sequence classifier
var teacher = new HiddenMarkovClassifierLearning <Independent>(
classifier,
// Train each model until the log-likelihood changes less than 0.0001
modelIndex => new BaumWelchLearning <Independent>(
classifier.Models[modelIndex])
{
Tolerance = 0.0001,
Iterations = 0,
}
);
// Train the sequence classifier using the algorithm
double logLikelihood = teacher.Run(sequences2, labels2);
Assert.AreEqual(-3.0493028798326081d, logLikelihood, 1e-10);
return(classifier);
}
public void SimpleGestureRecognitionTest()
{
// Let's say we would like to do a very simple mechanism for
// gesture recognition. In this example, we will be trying to
// create a classifier that can distinguish between the words
// "hello", "car", and "wardrobe".
// Let's say we decided to acquire some data, and we asked some
// people to perform those words in front of a Kinect camera, and,
// using Microsoft's SDK, we were able to captured the x and y
// coordinates of each hand while the word was being performed.
// Let's say we decided to represent our frames as:
//
// double[] frame = { leftHandX, leftHandY, rightHandX, rightHandY };
//
// Since we captured words, this means we captured sequences of
// frames as we described above. Let's write some of those as
// rough examples to explain how gesture recognition can be done:
double[][] hello =
{
new double[] { 1.0, 0.1, 0.0, 0.0 }, // let's say the word
new double[] { 0.0, 1.0, 0.1, 0.1 }, // hello took 6 frames
new double[] { 0.0, 1.0, 0.1, 0.1 }, // to be recorded.
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 0.1, 1.1 },
};
double[][] car =
{
new double[] { 0.0, 0.0, 0.0, 1.0 }, // the car word
new double[] { 0.1, 0.0, 1.0, 0.1 }, // took only 4.
new double[] { 0.0, 0.0, 0.1, 0.0 },
new double[] { 1.0, 0.0, 0.0, 0.0 },
};
double[][] wardrobe =
{
new double[] { 0.0, 0.0, 1.0, 0.0 }, // same for the
new double[] { 0.1, 0.0, 1.0, 0.1 }, // wardrobe word.
new double[] { 0.0, 0.1, 1.0, 0.0 },
new double[] { 0.1, 0.0, 1.0, 0.1 },
};
// Here, please note that a real-world example would involve *lots*
// of samples for each word. Here, we are considering just one from
// each class which is clearly sub-optimal and should _never_ be done
// on practice. For example purposes, however, please disregard this.
// Those are the words we have in our vocabulary:
//
double[][][] words = { hello, car, wardrobe };
// Now, let's associate integer labels with them. This is needed
// for the case where there are multiple samples for each word.
//
int[] labels = { 0, 1, 2 };
// We will create our classifiers assuming an independent
// Gaussian distribution for each component in our feature
// vectors (like assuming a Naive Bayes assumption).
Independent initial = new Independent
(
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1)
);
// Now, we can proceed and create our classifier.
//
int numberOfWords = 3; // we are trying to distinguish between 3 words
int numberOfStates = 5; // this value can be found by trial-and-error
var classifier = new HiddenMarkovClassifier <Independent>
(
classes: numberOfWords,
topology: new Forward(numberOfStates), // word classifiers should use a forward topology
initial: initial
);
// Create a new learning algorithm to train the sequence classifier
var teacher = new HiddenMarkovClassifierLearning <Independent>(classifier,
// Train each model until the log-likelihood changes less than 0.001
modelIndex => new BaumWelchLearning <Independent>(classifier.Models[modelIndex])
{
Tolerance = 0.001,
Iterations = 100,
// This is necessary so the code doesn't blow up when it realize
// there is only one sample per word class. But this could also be
// needed in normal situations as well.
//
FittingOptions = new IndependentOptions()
{
InnerOption = new NormalOptions()
{
Regularization = 1e-5
}
}
}
);
// Finally, we can run the learning algorithm!
double logLikelihood = teacher.Run(words, labels);
// At this point, the classifier should be successfully
// able to distinguish between our three word classes:
//
int c1 = classifier.Compute(hello);
int c2 = classifier.Compute(car);
int c3 = classifier.Compute(wardrobe);
Assert.AreEqual(0, c1);
Assert.AreEqual(1, c2);
Assert.AreEqual(2, c3);
}
public void CumulativeFunctionTest()
{
var p1 = new NormalDistribution(4.2, 1);
var p2 = new NormalDistribution(7.0, 2);
Independent<NormalDistribution> target = new Independent<NormalDistribution>(p1, p2);
double[] x;
double actual, expected;
x = new double[] { 4.2, 7.0 };
actual = target.DistributionFunction(x);
expected = p1.DistributionFunction(x[0]) * p2.DistributionFunction(x[1]);
Assert.AreEqual(expected, actual);
x = new double[] { 0.0, 0.0 };
actual = target.DistributionFunction(x);
expected = p1.DistributionFunction(x[0]) * p2.DistributionFunction(x[1]);
Assert.AreEqual(expected, actual);
x = new double[] { 7.0, 4.2 };
actual = target.DistributionFunction(x);
expected = p1.DistributionFunction(x[0]) * p2.DistributionFunction(x[1]);
Assert.AreEqual(expected, actual);
}
public void FitTest()
{
double[][] data =
{
new double[] { 1, 8 },
new double[] { 2, 6 },
new double[] { 5, 7 },
new double[] { 3, 9 },
};
var p1 = new NormalDistribution(0, 1);
var p2 = new NormalDistribution(0, 1);
Independent<NormalDistribution> target = new Independent<NormalDistribution>(p1, p2);
target.Fit(data);
Assert.AreEqual(2.75, target.Mean[0]);
Assert.AreEqual(7.50, target.Mean[1]);
Assert.AreEqual(2.91666, target.Variance[0], 1e-5);
Assert.AreEqual(1.66666, target.Variance[1], 1e-5);
Assert.AreEqual(target.Variance[0], target.Covariance[0, 0]);
Assert.AreEqual(target.Variance[1], target.Covariance[1, 1]);
Assert.AreEqual(0, target.Covariance[0, 1]);
Assert.AreEqual(0, target.Covariance[1, 0]);
}
public static HiddenMarkovClassifier<Independent> CreateModel2(out double[][][] sequences, out int[] labels)
{
sequences = new double[][][]
{
new double[][]
{
// This is the first sequence with label = 0
new double[] { 0, 1.1 },
new double[] { 1, 2.5 },
new double[] { 1, 3.4 },
new double[] { 1, 4.7 },
new double[] { 2, 5.8 },
},
new double[][]
{
// This is the second sequence with label = 1
new double[] { 2, 3.2 },
new double[] { 2, 2.6 },
new double[] { 1, 1.2 },
new double[] { 1, 0.8 },
new double[] { 0, 1.1 },
}
};
labels = new[] { 0, 1 };
// Create a Continuous density Hidden Markov Model Sequence Classifier
// to detect a multivariate sequence and the same sequence backwards.
var comp1 = new GeneralDiscreteDistribution(3);
var comp2 = new NormalDistribution(1);
var density = new Independent(comp1, comp2);
// Creates a sequence classifier containing 2 hidden Markov Models with 2 states
// and an underlying multivariate mixture of Normal distributions as density.
var classifier = new HiddenMarkovClassifier<Independent>(
2, new Ergodic(2), density);
// Configure the learning algorithms to train the sequence classifier
var teacher = new HiddenMarkovClassifierLearning<Independent>(
classifier,
// Train each model until the log-likelihood changes less than 0.0001
modelIndex => new BaumWelchLearning<Independent>(
classifier.Models[modelIndex])
{
Tolerance = 0.0001,
Iterations = 0,
}
);
// Train the sequence classifier using the algorithm
double logLikelihood = teacher.Run(sequences, labels);
return classifier;
}
public void LearnTest10_Independent()
{
// Let's say we have 2 meteorological sensors gathering data
// from different time periods of the day. Those periods are
// represented below:
double[][][] data =
{
new double[][] // first sequence (we just repeated the measurements
{ // once, so there is only one observation sequence)
new double[] { 1, 2 }, // Day 1, 15:00 pm
new double[] { 6, 7 }, // Day 1, 16:00 pm
new double[] { 2, 3 }, // Day 1, 17:00 pm
new double[] { 2, 2 }, // Day 1, 18:00 pm
new double[] { 9, 8 }, // Day 1, 19:00 pm
new double[] { 1, 0 }, // Day 1, 20:00 pm
new double[] { 1, 3 }, // Day 1, 21:00 pm
new double[] { 8, 9 }, // Day 1, 22:00 pm
new double[] { 3, 3 }, // Day 1, 23:00 pm
new double[] { 1, 3 }, // Day 2, 00:00 am
new double[] { 1, 1 }, // Day 2, 01:00 am
}
};
// Let's assume those sensors are unrelated (for simplicity). As
// such, let's assume the data gathered from the sensors may reside
// into circular centroids denoting each state the underlying system
// might be in.
NormalDistribution[] initial_components =
{
new NormalDistribution(), // initial value for the first variable's distribution
new NormalDistribution() // initial value for the second variable's distribution
};
// Specify a initial independent normal distribution for the samples.
var density = new Independent<NormalDistribution>(initial_components);
// Creates a continuous hidden Markov Model with two states organized in an Ergodic
// topology and an underlying independent Normal distribution as probability density.
var model = new HiddenMarkovModel<Independent<NormalDistribution>>(new Ergodic(2), density);
// Configure the learning algorithms to train the sequence classifier until the
// difference in the average log-likelihood changes only by as little as 0.0001
var teacher = new BaumWelchLearning<Independent<NormalDistribution>>(model)
{
Tolerance = 0.0001,
Iterations = 0,
};
// Fit the model
double error = teacher.Run(data);
// Get the hidden state associated with each observation
//
double logLikelihood; // log-likelihood of the Viterbi path
int[] hidden_states = model.Decode(data[0], out logLikelihood);
Assert.AreEqual(-33.978800850637882, error);
Assert.AreEqual(-33.9788008509802, logLikelihood);
Assert.AreEqual(11, hidden_states.Length);
}
public void SingleDependencyBeforeUpdateIsAsSmallAsPossible()
{
GC.Collect();
long start = GC.GetTotalMemory(true);
Independent<int> newIndependent = new Independent<int>();
Dependent<int> newDependent = new Dependent<int>(() => newIndependent);
newIndependent.Value = 42;
long end = GC.GetTotalMemory(true);
// Started at 336.
// Making Precedent a base class: 312.
// Removing Gain/LoseDependent events: 288.
// Making IsUpToDate no longer a precident: 248.
// Custom linked list implementation for dependents: 200.
// Custom linked list implementation for precedents: 160.
// Other optimizations: 144.
// Added WeakReferenceToSelf: 148.
// Removed WeakReferenceToSelf: 124.
Assert.AreEqual(124 + IndependentPlatformOffset, end - start);
int value = newDependent;
Assert.AreEqual(42, value);
}
/// <summary>
/// Initializes a new instance of the <see cref="MetropolisHasting"/> algorithm.
/// </summary>
///
/// <param name="dimensions">The number of dimensions in each observation.</param>
/// <param name="logDensity">A function specifying the log probability density of the distribution to be sampled.</param>
/// <param name="proposal">The proposal distribution that is used to generate new parameter samples to be explored.</param>
///
public MetropolisHasting(int dimensions, Func <double[], double> logDensity, Independent <NormalDistribution> proposal)
: base(dimensions, logDensity, proposal)
{
}
public void FittingNoOptionsTest()
{
double[][] data1 =
{
new double[] { 0, 8 },
new double[] { 0, 6 },
new double[] { 0, 7 },
new double[] { 0, 9 },
};
double[][] data2 =
{
new double[] { 8, 0 },
new double[] { 6, 0 },
new double[] { 7, 0 },
new double[] { 9, 0 },
};
var p1 = new NormalDistribution(0, 1);
var p2 = new NormalDistribution(0, 1);
bool thrown1 = false;
bool thrown2 = false;
try
{
var target = new Independent<NormalDistribution>(p1, p2);
target.Fit(data1);
}
catch (ArgumentException)
{
thrown1 = true;
}
try
{
var target = new Independent<NormalDistribution>(p1, p2);
target.Fit(data2);
}
catch (ArgumentException)
{
thrown2 = true;
}
Assert.IsTrue(thrown1);
Assert.IsTrue(thrown2);
}
public static HiddenMarkovClassifier<Independent> CreateModel3(out double[][][] sequences2, out int[] labels2)
{
sequences2 = new double[][][]
{
new double[][]
{
// This is the first sequence with label = 0
new double[] { 1, 1.12, 2.41, 1.17, 9.3 },
new double[] { 1, 2.54, 1.45, 0.16, 4.5 },
new double[] { 1, 3.46, 2.63, 1.15, 9.2 },
new double[] { 1, 4.73, 0.41, 1.54, 5.5 },
new double[] { 2, 5.81, 2.42, 1.13, 9.1 },
},
new double[][]
{
// This is the first sequence with label = 0
new double[] { 0, 1.49, 2.48, 1.18, 9.37 },
new double[] { 1, 2.18, 1.44, 2.19, 1.56 },
new double[] { 1, 3.77, 2.62, 1.10, 9.25 },
new double[] { 2, 4.76, 5.44, 3.58, 5.54 },
new double[] { 2, 5.85, 2.46, 1.16, 5.13 },
new double[] { 2, 4.84, 5.44, 3.54, 5.52 },
new double[] { 2, 5.83, 3.41, 1.22, 5.11 },
},
new double[][]
{
// This is the first sequence with label = 0
new double[] { 2, 1.11, 2.41, 1.12, 2.31 },
new double[] { 1, 2.52, 3.73, 0.12, 4.50 },
new double[] { 1, 3.43, 2.61, 1.24, 9.29 },
new double[] { 1, 4.74, 2.42, 2.55, 6.57 },
new double[] { 2, 5.85, 2.43, 1.16, 9.16 },
},
new double[][]
{
// This is the second sequence with label = 1
new double[] { 0, 1.26, 5.44, 1.56, 9.55 },
new double[] { 2, 2.67, 5.45, 4.27, 1.54 },
new double[] { 1, 1.28, 3.46, 2.18, 4.13 },
new double[] { 1, 5.89, 2.57, 1.79, 5.02 },
new double[] { 0, 1.40, 2.48, 2.10, 6.41 },
},
new double[][]
{
// This is the second sequence with label = 1
new double[] { 2, 3.21, 2.49, 1.54, 9.17 },
new double[] { 2, 2.62, 5.40, 4.25, 1.54 },
new double[] { 1, 1.53, 6.49, 2.17, 4.52 },
new double[] { 1, 2.84, 2.58, 1.73, 6.04 },
new double[] { 1, 1.45, 2.47, 2.28, 5.42 },
new double[] { 1, 1.46, 2.46, 2.35, 5.41 },
},
new double[][]
{
// This is the second sequence with label = 1
new double[] { 1, 5.27, 5.45, 1.4, 9.5 },
new double[] { 2, 2.68, 2.54, 3.2, 2.2 },
new double[] { 1, 2.89, 3.83, 2.6, 4.1 },
new double[] { 1, 1.80, 1.32, 1.2, 4.2 },
new double[] { 0, 1.41, 2.41, 2.1, 6.4 },
}
};
labels2 = new[] { 0, 0, 0, 1, 1, 1 };
// Create a Continuous density Hidden Markov Model Sequence Classifier
// to detect a multivariate sequence and the same sequence backwards.
var comp1 = new GeneralDiscreteDistribution(3);
var comp2 = new NormalDistribution(1);
var comp3 = new NormalDistribution(2);
var comp4 = new NormalDistribution(3);
var comp5 = new NormalDistribution(4);
var density = new Independent(comp1, comp2, comp3, comp4, comp5);
// Creates a sequence classifier containing 2 hidden Markov Models with 2 states
// and an underlying multivariate mixture of Normal distributions as density.
var classifier = new HiddenMarkovClassifier<Independent>(
2, new Forward(5), density);
// Configure the learning algorithms to train the sequence classifier
var teacher = new HiddenMarkovClassifierLearning<Independent>(
classifier,
// Train each model until the log-likelihood changes less than 0.0001
modelIndex => new BaumWelchLearning<Independent>(
classifier.Models[modelIndex])
{
Tolerance = 0.0001,
Iterations = 0,
}
);
// Train the sequence classifier using the algorithm
double logLikelihood = teacher.Run(sequences2, labels2);
return classifier;
}
public void SimpleGestureRecognitionTest()
{
// Let's say we would like to do a very simple mechanism for
// gesture recognition. In this example, we will be trying to
// create a classifier that can distinguish between the words
// "hello", "car", and "wardrobe".
// Let's say we decided to acquire some data, and we asked some
// people to perform those words in front of a Kinect camera, and,
// using Microsoft's SDK, we were able to captured the x and y
// coordinates of each hand while the word was being performed.
// Let's say we decided to represent our frames as:
//
// double[] frame = { leftHandX, leftHandY, rightHandX, rightHandY };
//
// Since we captured words, this means we captured sequences of
// frames as we described above. Let's write some of those as
// rough examples to explain how gesture recognition can be done:
double[][] hello =
{
new double[] { 1.0, 0.1, 0.0, 0.0 }, // let's say the word
new double[] { 0.0, 1.0, 0.1, 0.1 }, // hello took 6 frames
new double[] { 0.0, 1.0, 0.1, 0.1 }, // to be recorded.
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 0.1, 1.1 },
};
double[][] car =
{
new double[] { 0.0, 0.0, 0.0, 1.0 }, // the car word
new double[] { 0.1, 0.0, 1.0, 0.1 }, // took only 4.
new double[] { 0.0, 0.0, 0.1, 0.0 },
new double[] { 1.0, 0.0, 0.0, 0.0 },
};
double[][] wardrobe =
{
new double[] { 0.0, 0.0, 1.0, 0.0 }, // same for the
new double[] { 0.1, 0.0, 1.0, 0.1 }, // wardrobe word.
new double[] { 0.0, 0.1, 1.0, 0.0 },
new double[] { 0.1, 0.0, 1.0, 0.1 },
};
// Here, please note that a real-world example would involve *lots*
// of samples for each word. Here, we are considering just one from
// each class which is clearly sub-optimal and should _never_ be done
// on practice. For example purposes, however, please disregard this.
// Those are the words we have in our vocabulary:
//
double[][][] words = { hello, car, wardrobe };
// Now, let's associate integer labels with them. This is needed
// for the case where there are multiple samples for each word.
//
int[] labels = { 0, 1, 2 };
// We will create our classifiers assuming an independent
// Gaussian distribution for each component in our feature
// vectors (like assuming a Naive Bayes assumption).
var initial = new Independent <NormalDistribution>
(
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1)
);
// Now, we can proceed and create our classifier.
//
int numberOfWords = 3; // we are trying to distinguish between 3 words
int numberOfStates = 5; // this value can be found by trial-and-error
var hmm = new HiddenMarkovClassifier <Independent <NormalDistribution> >
(
classes: numberOfWords,
topology: new Forward(numberOfStates), // word classifiers should use a forward topology
initial: initial
);
// Create a new learning algorithm to train the sequence classifier
var teacher = new HiddenMarkovClassifierLearning <Independent <NormalDistribution> >(hmm,
// Train each model until the log-likelihood changes less than 0.001
modelIndex => new BaumWelchLearning <Independent <NormalDistribution> >(hmm.Models[modelIndex])
{
Tolerance = 0.001,
Iterations = 100,
// This is necessary so the code doesn't blow up when it realize
// there is only one sample per word class. But this could also be
// needed in normal situations as well.
//
FittingOptions = new IndependentOptions()
{
InnerOption = new NormalOptions()
{
Regularization = 1e-5
}
}
}
);
// Finally, we can run the learning algorithm!
double logLikelihood = teacher.Run(words, labels);
// At this point, the classifier should be successfully
// able to distinguish between our three word classes:
//
int tc1 = hmm.Compute(hello);
int tc2 = hmm.Compute(car);
int tc3 = hmm.Compute(wardrobe);
Assert.AreEqual(0, tc1);
Assert.AreEqual(1, tc2);
Assert.AreEqual(2, tc3);
// Now, we can use the Markov classifier to initialize a HCRF
var function = new MarkovMultivariateFunction(hmm);
var hcrf = new HiddenConditionalRandomField <double[]>(function);
// We can check that both are equivalent, although they have
// formulations that can be learned with different methods
//
for (int i = 0; i < words.Length; i++)
{
// Should be the same
int expected = hmm.Compute(words[i]);
int actual = hcrf.Compute(words[i]);
// Should be the same
double h0 = hmm.LogLikelihood(words[i], 0);
double c0 = hcrf.LogLikelihood(words[i], 0);
double h1 = hmm.LogLikelihood(words[i], 1);
double c1 = hcrf.LogLikelihood(words[i], 1);
double h2 = hmm.LogLikelihood(words[i], 2);
double c2 = hcrf.LogLikelihood(words[i], 2);
Assert.AreEqual(expected, actual);
Assert.AreEqual(h0, c0, 1e-10);
Assert.IsTrue(h1.IsRelativelyEqual(c1, 1e-10));
Assert.IsTrue(h2.IsRelativelyEqual(c2, 1e-10));
Assert.IsFalse(double.IsNaN(c0));
Assert.IsFalse(double.IsNaN(c1));
Assert.IsFalse(double.IsNaN(c2));
}
// Now we can learn the HCRF using one of the best learning
// algorithms available, Resilient Backpropagation learning:
// Create a learning algorithm
var rprop = new HiddenResilientGradientLearning <double[]>(hcrf)
{
Iterations = 50,
Tolerance = 1e-5
};
// Run the algorithm and learn the models
double error = rprop.Run(words, labels);
// At this point, the HCRF should be successfully
// able to distinguish between our three word classes:
//
int hc1 = hcrf.Compute(hello);
int hc2 = hcrf.Compute(car);
int hc3 = hcrf.Compute(wardrobe);
Assert.AreEqual(0, hc1);
Assert.AreEqual(1, hc2);
Assert.AreEqual(2, hc3);
}
public static HiddenMarkovClassifier<Independent<NormalDistribution>> CreateModel4(out double[][][] words, out int[] labels, bool usePriors)
{
double[][] hello =
{
new double[] { 1.0, 0.1, 0.0, 0.0 }, // let's say the word
new double[] { 0.0, 1.0, 0.1, 0.1 }, // hello took 6 frames
new double[] { 0.0, 1.0, 0.1, 0.1 }, // to be recorded.
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 0.1, 1.1 },
};
double[][] car =
{
new double[] { 0.0, 0.0, 0.0, 1.0 }, // the car word
new double[] { 0.1, 0.0, 1.0, 0.1 }, // took only 4.
new double[] { 0.0, 0.0, 0.1, 0.0 },
new double[] { 1.0, 0.0, 0.0, 0.0 },
};
double[][] wardrobe =
{
new double[] { 0.0, 0.0, 1.0, 0.0 }, // same for the
new double[] { 0.1, 0.0, 1.0, 0.1 }, // wardrobe word.
new double[] { 0.0, 0.1, 1.0, 0.0 },
new double[] { 0.1, 0.0, 1.0, 0.1 },
};
double[][] wardrobe2 =
{
new double[] { 0.0, 0.0, 1.0, 0.0 }, // same for the
new double[] { 0.2, 0.0, 1.0, 0.1 }, // wardrobe word.
new double[] { 0.0, 0.1, 1.0, 0.0 },
new double[] { 0.1, 0.0, 1.0, 0.2 },
};
words = new double[][][] { hello, car, wardrobe, wardrobe2 };
labels = new [] { 0, 1, 2, 2 };
var initial = new Independent<NormalDistribution>
(
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1)
);
int numberOfWords = 3;
int numberOfStates = 5;
var classifier = new HiddenMarkovClassifier<Independent<NormalDistribution>>
(
classes: numberOfWords,
topology: new Forward(numberOfStates),
initial: initial
);
var teacher = new HiddenMarkovClassifierLearning<Independent<NormalDistribution>>(classifier,
modelIndex => new BaumWelchLearning<Independent<NormalDistribution>>(classifier.Models[modelIndex])
{
Tolerance = 0.001,
Iterations = 100,
FittingOptions = new IndependentOptions()
{
InnerOption = new NormalOptions() { Regularization = 1e-5 }
}
}
);
if (usePriors)
teacher.Empirical = true;
double logLikelihood = teacher.Run(words, labels);
return classifier;
}
public void SaveLoadTest()
{
double[][] hello =
{
new double[] { 1.0, 0.1, 0.0, 0.0 }, // let's say the word
new double[] { 0.0, 1.0, 0.1, 0.1 }, // hello took 6 frames
new double[] { 0.0, 1.0, 0.1, 0.1 }, // to be recorded.
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 0.1, 1.1 },
};
double[][] car =
{
new double[] { 0.0, 0.0, 0.0, 1.0 }, // the car word
new double[] { 0.1, 0.0, 1.0, 0.1 }, // took only 4.
new double[] { 0.0, 0.0, 0.1, 0.0 },
new double[] { 1.0, 0.0, 0.0, 0.0 },
};
double[][] wardrobe =
{
new double[] { 0.0, 0.0, 1.0, 0.0 }, // same for the
new double[] { 0.1, 0.0, 1.0, 0.1 }, // wardrobe word.
new double[] { 0.0, 0.1, 1.0, 0.0 },
new double[] { 0.1, 0.0, 1.0, 0.1 },
};
double[][][] words = { hello, car, wardrobe };
int[] labels = { 0, 1, 2 };
var initial = new Independent
(
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1)
);
int numberOfWords = 3;
int numberOfStates = 5;
var classifier = new HiddenMarkovClassifier <Independent>
(
classes: numberOfWords,
topology: new Forward(numberOfStates),
initial: initial
);
var teacher = new HiddenMarkovClassifierLearning <Independent>(classifier,
modelIndex => new BaumWelchLearning <Independent>(classifier.Models[modelIndex])
{
Tolerance = 0.001,
Iterations = 100,
FittingOptions = new IndependentOptions()
{
InnerOption = new NormalOptions()
{
Regularization = 1e-5
}
}
}
);
double logLikelihood = teacher.Run(words, labels);
var function = new MarkovMultivariateFunction(classifier);
var hcrf = new HiddenConditionalRandomField <double[]>(function);
MemoryStream stream = new MemoryStream();
hcrf.Save(stream);
stream.Seek(0, SeekOrigin.Begin);
var target = HiddenConditionalRandomField <double[]> .Load(stream);
Assert.AreEqual(hcrf.Function.Factors.Length, target.Function.Factors.Length);
for (int i = 0; i < hcrf.Function.Factors.Length; i++)
{
var e = hcrf.Function.Factors[i];
var a = target.Function.Factors[i];
Assert.AreEqual(e.Index, target.Function.Factors[i].Index);
Assert.AreEqual(e.States, target.Function.Factors[i].States);
Assert.AreEqual(e.EdgeParameters.Count, a.EdgeParameters.Count);
Assert.AreEqual(e.EdgeParameters.Offset, a.EdgeParameters.Offset);
Assert.AreEqual(e.FactorParameters.Count, a.FactorParameters.Count);
Assert.AreEqual(e.FactorParameters.Offset, a.FactorParameters.Offset);
Assert.AreEqual(e.OutputParameters.Count, a.OutputParameters.Count);
Assert.AreEqual(e.OutputParameters.Offset, a.OutputParameters.Offset);
Assert.AreEqual(e.StateParameters.Count, a.StateParameters.Count);
Assert.AreEqual(e.StateParameters.Offset, a.StateParameters.Offset);
Assert.AreEqual(target.Function, a.Owner);
Assert.AreEqual(hcrf.Function, e.Owner);
}
Assert.AreEqual(hcrf.Function.Features.Length, target.Function.Features.Length);
for (int i = 0; i < hcrf.Function.Factors.Length; i++)
{
Assert.AreEqual(hcrf.Function.Features[i].GetType(), target.Function.Features[i].GetType());
}
Assert.AreEqual(hcrf.Function.Outputs, target.Function.Outputs);
for (int i = 0; i < hcrf.Function.Weights.Length; i++)
{
Assert.AreEqual(hcrf.Function.Weights[i], target.Function.Weights[i]);
}
}
public static HiddenMarkovClassifier<Independent> CreateModel3()
{
// Create a Continuous density Hidden Markov Model Sequence Classifier
// to detect a multivariate sequence and the same sequence backwards.
var comp1 = new GeneralDiscreteDistribution(3);
var comp2 = new NormalDistribution(1);
var comp3 = new NormalDistribution(2);
var comp4 = new NormalDistribution(3);
var comp5 = new NormalDistribution(4);
var density = new Independent(comp1, comp2, comp3, comp4, comp5);
// Creates a sequence classifier containing 2 hidden Markov Models with 2 states
// and an underlying multivariate mixture of Normal distributions as density.
var classifier = new HiddenMarkovClassifier<Independent>(
2, new Forward(5), density);
// Configure the learning algorithms to train the sequence classifier
var teacher = new HiddenMarkovClassifierLearning<Independent>(
classifier,
// Train each model until the log-likelihood changes less than 0.0001
modelIndex => new BaumWelchLearning<Independent>(
classifier.Models[modelIndex])
{
Tolerance = 0.0001,
Iterations = 0,
}
);
// Train the sequence classifier using the algorithm
double logLikelihood = teacher.Run(sequences2, labels2);
return classifier;
}
public void IndependentModelExample()
{
double[][] samples =
{
new double[] { 0, 1 },
new double[] { 1, 2 },
new double[] { 1, 1 },
new double[] { 0, 7 },
new double[] { 1, 1 },
new double[] { 6, 2 },
new double[] { 6, 5 },
new double[] { 5, 1 },
new double[] { 7, 1 },
new double[] { 5, 1 }
};
// Create a new 2D independent Normal distribution
var independent = new Independent<NormalDistribution>(
new NormalDistribution(), new NormalDistribution());
// Estimate it!
independent.Fit(samples);
var a = independent.Components[0].ToString("N2"); // N(x1; μ = 3.20, σ² = 7.96)
var b = independent.Components[1].ToString("N2"); // N(x2; μ = 2.20, σ² = 4.40)
Assert.AreEqual("N(x; μ = 3.20, σ² = 7.96)", a);
Assert.AreEqual("N(x; μ = 2.20, σ² = 4.40)", b);
}
public void IndependentIsAsSmallAsPossible()
{
GC.Collect();
long start = GC.GetTotalMemory(true);
Independent<int> newIndependent = new Independent<int>();
newIndependent.Value = 42;
long end = GC.GetTotalMemory(true);
// Started at 92.
// Making Precedent a base class: 80.
// Removing Gain/LoseDependent events: 72.
// Custom linked list implementation for dependents: 48.
// Other optimizations: 40.
// Removed WeakReferenceToSelf: 20.
Assert.AreEqual(20 + IndependentPlatformOffset, end - start);
int value = newIndependent;
Assert.AreEqual(42, value);
}
public void NumberOfFeaturesTest()
{
Independent initial = new Independent(
new GeneralDiscreteDistribution(46), // output,
new NormalDistribution(0, 1), // headAngle,
new NormalDistribution(0, 1), // handAngle,
new NormalDistribution(0, 1), // relHandX,
new NormalDistribution(0, 1) // relHandY,
);
var model = new HiddenMarkovClassifier<Independent>(
classes: 13, topology: new Forward(5), initial: initial);
var function = new MarkovMultivariateFunction(model);
var field = new HiddenConditionalRandomField<double[]>(function);
int discreteDistributions = 1;
int continuousDistributions = 4;
int symbols = 46;
int states = 5;
int classes = 13;
int expected = classes *
(1 + states + states * states +
states * discreteDistributions * symbols +
states * continuousDistributions * 3);
Assert.AreEqual(expected, field.Function.Features.Length);
Assert.AreEqual(expected, field.Function.Weights.Length);
Assert.AreEqual(4173, expected);
}
/// <summary>
/// Serves as a hash function for a particular type. <see cref="M:System.Object.GetHashCode"></see> is suitable for use in hashing algorithms and data structures like a hash table.
/// </summary>
/// <returns>
/// A hash code for the current <see cref="T:System.Object"></see>.
/// </returns>
public override int GetHashCode()
{
return((Dependent == null ? 0 : Dependent.GetHashCode()) ^ (Independent == null ? 0 : Independent.GetHashCode()));
}
public void SaveLoadTest()
{
double[][] hello =
{
new double[] { 1.0, 0.1, 0.0, 0.0 }, // let's say the word
new double[] { 0.0, 1.0, 0.1, 0.1 }, // hello took 6 frames
new double[] { 0.0, 1.0, 0.1, 0.1 }, // to be recorded.
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 0.1, 1.1 },
};
double[][] car =
{
new double[] { 0.0, 0.0, 0.0, 1.0 }, // the car word
new double[] { 0.1, 0.0, 1.0, 0.1 }, // took only 4.
new double[] { 0.0, 0.0, 0.1, 0.0 },
new double[] { 1.0, 0.0, 0.0, 0.0 },
};
double[][] wardrobe =
{
new double[] { 0.0, 0.0, 1.0, 0.0 }, // same for the
new double[] { 0.1, 0.0, 1.0, 0.1 }, // wardrobe word.
new double[] { 0.0, 0.1, 1.0, 0.0 },
new double[] { 0.1, 0.0, 1.0, 0.1 },
};
double[][][] words = { hello, car, wardrobe };
int[] labels = { 0, 1, 2 };
var initial = new Independent
(
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1)
);
int numberOfWords = 3;
int numberOfStates = 5;
var classifier = new HiddenMarkovClassifier<Independent>
(
classes: numberOfWords,
topology: new Forward(numberOfStates),
initial: initial
);
var teacher = new HiddenMarkovClassifierLearning<Independent>(classifier,
modelIndex => new BaumWelchLearning<Independent>(classifier.Models[modelIndex])
{
Tolerance = 0.001,
Iterations = 100,
FittingOptions = new IndependentOptions()
{
InnerOption = new NormalOptions() { Regularization = 1e-5 }
}
}
);
double logLikelihood = teacher.Run(words, labels);
var function = new MarkovMultivariateFunction(classifier);
var hcrf = new HiddenConditionalRandomField<double[]>(function);
MemoryStream stream = new MemoryStream();
hcrf.Save(stream);
stream.Seek(0, SeekOrigin.Begin);
var target = HiddenConditionalRandomField<double[]>.Load(stream);
Assert.AreEqual(hcrf.Function.Factors.Length, target.Function.Factors.Length);
for (int i = 0; i < hcrf.Function.Factors.Length; i++)
{
var e = hcrf.Function.Factors[i];
var a = target.Function.Factors[i];
Assert.AreEqual(e.Index, target.Function.Factors[i].Index);
Assert.AreEqual(e.States, target.Function.Factors[i].States);
Assert.AreEqual(e.EdgeParameters.Count, a.EdgeParameters.Count);
Assert.AreEqual(e.EdgeParameters.Offset, a.EdgeParameters.Offset);
Assert.AreEqual(e.FactorParameters.Count, a.FactorParameters.Count);
Assert.AreEqual(e.FactorParameters.Offset, a.FactorParameters.Offset);
Assert.AreEqual(e.OutputParameters.Count, a.OutputParameters.Count);
Assert.AreEqual(e.OutputParameters.Offset, a.OutputParameters.Offset);
Assert.AreEqual(e.StateParameters.Count, a.StateParameters.Count);
Assert.AreEqual(e.StateParameters.Offset, a.StateParameters.Offset);
Assert.AreEqual(target.Function, a.Owner);
Assert.AreEqual(hcrf.Function, e.Owner);
}
Assert.AreEqual(hcrf.Function.Features.Length, target.Function.Features.Length);
for (int i = 0; i < hcrf.Function.Factors.Length; i++)
Assert.AreEqual(hcrf.Function.Features[i].GetType(), target.Function.Features[i].GetType());
Assert.AreEqual(hcrf.Function.Outputs, target.Function.Outputs);
for (int i = 0; i < hcrf.Function.Weights.Length; i++)
Assert.AreEqual(hcrf.Function.Weights[i], target.Function.Weights[i]);
}
public void SimpleGestureRecognitionTest()
{
// Let's say we would like to do a very simple mechanism for
// gesture recognition. In this example, we will be trying to
// create a classifier that can distinguish between the words
// "hello", "car", and "wardrobe".
// Let's say we decided to acquire some data, and we asked some
// people to perform those words in front of a Kinect camera, and,
// using Microsoft's SDK, we were able to captured the x and y
// coordinates of each hand while the word was being performed.
// Let's say we decided to represent our frames as:
//
// double[] frame = { leftHandX, leftHandY, rightHandX, rightHandY };
//
// Since we captured words, this means we captured sequences of
// frames as we described above. Let's write some of those as
// rough examples to explain how gesture recognition can be done:
double[][] hello =
{
new double[] { 1.0, 0.1, 0.0, 0.0 }, // let's say the word
new double[] { 0.0, 1.0, 0.1, 0.1 }, // hello took 6 frames
new double[] { 0.0, 1.0, 0.1, 0.1 }, // to be recorded.
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 0.1, 1.1 },
};
double[][] car =
{
new double[] { 0.0, 0.0, 0.0, 1.0 }, // the car word
new double[] { 0.1, 0.0, 1.0, 0.1 }, // took only 4.
new double[] { 0.0, 0.0, 0.1, 0.0 },
new double[] { 1.0, 0.0, 0.0, 0.0 },
};
double[][] wardrobe =
{
new double[] { 0.0, 0.0, 1.0, 0.0 }, // same for the
new double[] { 0.1, 0.0, 1.0, 0.1 }, // wardrobe word.
new double[] { 0.0, 0.1, 1.0, 0.0 },
new double[] { 0.1, 0.0, 1.0, 0.1 },
};
// Here, please note that a real-world example would involve *lots*
// of samples for each word. Here, we are considering just one from
// each class which is clearly sub-optimal and should _never_ be done
// on practice. For example purposes, however, please disregard this.
// Those are the words we have in our vocabulary:
//
double[][][] words = { hello, car, wardrobe };
// Now, let's associate integer labels with them. This is needed
// for the case where there are multiple samples for each word.
//
int[] labels = { 0, 1, 2 };
// We will create our classifiers assuming an independent
// Gaussian distribution for each component in our feature
// vectors (like assuming a Naive Bayes assumption).
Independent initial = new Independent
(
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1)
);
// Now, we can proceed and create our classifier.
//
int numberOfWords = 3; // we are trying to distinguish between 3 words
int numberOfStates = 5; // this value can be found by trial-and-error
var classifier = new HiddenMarkovClassifier<Independent>
(
classes: numberOfWords,
topology: new Forward(numberOfStates), // word classifiers should use a forward topology
initial: initial
);
// Create a new learning algorithm to train the sequence classifier
var teacher = new HiddenMarkovClassifierLearning<Independent>(classifier,
// Train each model until the log-likelihood changes less than 0.001
modelIndex => new BaumWelchLearning<Independent>(classifier.Models[modelIndex])
{
Tolerance = 0.001,
Iterations = 100,
// This is necessary so the code doesn't blow up when it realize
// there is only one sample per word class. But this could also be
// needed in normal situations as well.
//
FittingOptions = new IndependentOptions()
{
InnerOption = new NormalOptions() { Regularization = 1e-5 }
}
}
);
// Finally, we can run the learning algorithm!
double logLikelihood = teacher.Run(words, labels);
// At this point, the classifier should be successfully
// able to distinguish between our three word classes:
//
int c1 = classifier.Compute(hello);
int c2 = classifier.Compute(car);
int c3 = classifier.Compute(wardrobe);
Assert.AreEqual(0, c1);
Assert.AreEqual(1, c2);
Assert.AreEqual(2, c3);
}
public Dependent(Independent independent)
{
_Independent = independent;
}
public void SimpleGestureRecognitionTest()
{
// Let's say we would like to do a very simple mechanism for
// gesture recognition. In this example, we will be trying to
// create a classifier that can distinguish between the words
// "hello", "car", and "wardrobe".
// Let's say we decided to acquire some data, and we asked some
// people to perform those words in front of a Kinect camera, and,
// using Microsoft's SDK, we were able to captured the x and y
// coordinates of each hand while the word was being performed.
// Let's say we decided to represent our frames as:
//
// double[] frame = { leftHandX, leftHandY, rightHandX, rightHandY };
//
// Since we captured words, this means we captured sequences of
// frames as we described above. Let's write some of those as
// rough examples to explain how gesture recognition can be done:
double[][] hello =
{
new double[] { 1.0, 0.1, 0.0, 0.0 }, // let's say the word
new double[] { 0.0, 1.0, 0.1, 0.1 }, // hello took 6 frames
new double[] { 0.0, 1.0, 0.1, 0.1 }, // to be recorded.
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 1.0, 0.0 },
new double[] { 0.0, 0.0, 0.1, 1.1 },
};
double[][] car =
{
new double[] { 0.0, 0.0, 0.0, 1.0 }, // the car word
new double[] { 0.1, 0.0, 1.0, 0.1 }, // took only 4.
new double[] { 0.0, 0.0, 0.1, 0.0 },
new double[] { 1.0, 0.0, 0.0, 0.0 },
};
double[][] wardrobe =
{
new double[] { 0.0, 0.0, 1.0, 0.0 }, // same for the
new double[] { 0.1, 0.0, 1.0, 0.1 }, // wardrobe word.
new double[] { 0.0, 0.1, 1.0, 0.0 },
new double[] { 0.1, 0.0, 1.0, 0.1 },
};
// Here, please note that a real-world example would involve *lots*
// of samples for each word. Here, we are considering just one from
// each class which is clearly sub-optimal and should _never_ be done
// on practice. For example purposes, however, please disregard this.
// Those are the words we have in our vocabulary:
//
double[][][] words = { hello, car, wardrobe };
// Now, let's associate integer labels with them. This is needed
// for the case where there are multiple samples for each word.
//
int[] labels = { 0, 1, 2 };
// We will create our classifiers assuming an independent
// Gaussian distribution for each component in our feature
// vectors (like assuming a Naive Bayes assumption).
var initial = new Independent<NormalDistribution>
(
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1),
new NormalDistribution(0, 1)
);
// Now, we can proceed and create our classifier.
//
int numberOfWords = 3; // we are trying to distinguish between 3 words
int numberOfStates = 5; // this value can be found by trial-and-error
var hmm = new HiddenMarkovClassifier<Independent<NormalDistribution>>
(
classes: numberOfWords,
topology: new Forward(numberOfStates), // word classifiers should use a forward topology
initial: initial
);
// Create a new learning algorithm to train the sequence classifier
var teacher = new HiddenMarkovClassifierLearning<Independent<NormalDistribution>>(hmm,
// Train each model until the log-likelihood changes less than 0.001
modelIndex => new BaumWelchLearning<Independent<NormalDistribution>>(hmm.Models[modelIndex])
{
Tolerance = 0.001,
Iterations = 100,
// This is necessary so the code doesn't blow up when it realize
// there is only one sample per word class. But this could also be
// needed in normal situations as well.
//
FittingOptions = new IndependentOptions()
{
InnerOption = new NormalOptions() { Regularization = 1e-5 }
}
}
);
// Finally, we can run the learning algorithm!
double logLikelihood = teacher.Run(words, labels);
// At this point, the classifier should be successfully
// able to distinguish between our three word classes:
//
int tc1 = hmm.Compute(hello);
int tc2 = hmm.Compute(car);
int tc3 = hmm.Compute(wardrobe);
Assert.AreEqual(0, tc1);
Assert.AreEqual(1, tc2);
Assert.AreEqual(2, tc3);
// Now, we can use the Markov classifier to initialize a HCRF
var function = new MarkovMultivariateFunction(hmm);
var hcrf = new HiddenConditionalRandomField<double[]>(function);
// We can check that both are equivalent, although they have
// formulations that can be learned with different methods
//
for (int i = 0; i < words.Length; i++)
{
// Should be the same
int expected = hmm.Compute(words[i]);
int actual = hcrf.Compute(words[i]);
// Should be the same
double h0 = hmm.LogLikelihood(words[i], 0);
double c0 = hcrf.LogLikelihood(words[i], 0);
double h1 = hmm.LogLikelihood(words[i], 1);
double c1 = hcrf.LogLikelihood(words[i], 1);
double h2 = hmm.LogLikelihood(words[i], 2);
double c2 = hcrf.LogLikelihood(words[i], 2);
Assert.AreEqual(expected, actual);
Assert.AreEqual(h0, c0, 1e-10);
Assert.IsTrue(h1.IsRelativelyEqual(c1, 1e-10));
Assert.IsTrue(h2.IsRelativelyEqual(c2, 1e-10));
Assert.IsFalse(double.IsNaN(c0));
Assert.IsFalse(double.IsNaN(c1));
Assert.IsFalse(double.IsNaN(c2));
}
// Now we can learn the HCRF using one of the best learning
// algorithms available, Resilient Backpropagation learning:
// Create a learning algorithm
var rprop = new HiddenResilientGradientLearning<double[]>(hcrf)
{
Iterations = 50,
Tolerance = 1e-5
};
// Run the algorithm and learn the models
double error = rprop.Run(words, labels);
// At this point, the HCRF should be successfully
// able to distinguish between our three word classes:
//
int hc1 = hcrf.Compute(hello);
int hc2 = hcrf.Compute(car);
int hc3 = hcrf.Compute(wardrobe);
Assert.AreEqual(0, hc1);
Assert.AreEqual(1, hc2);
Assert.AreEqual(2, hc3);
}