/// <summary>
/// Constructs an instance from number of components and number of features
/// </summary>
/// <param name="nClass">Number of classes</param>
/// <param name="nFeatures">Number of features</param>
/// <param name="noisePrec">Noise precision</param>
public BPM(int nClass, int nFeatures, double noisePrec)
{
this.nClass = nClass;
this.nFeatures = nFeatures;
NoisePrec = noisePrec;
trainModel = SpecifyTrainModel("_train");
testModel = SpecifyTestModel("_test");
}
/// <summary>
/// Specifies the test model
/// </summary>
/// <param name="s">The name of the test model</param>
/// <returns>A <see cref="BPMVarsForTest"/> instance</returns>
private BPMVarsForTest SpecifyTestModel(string s)
{
// The number of test items - this will be set by the calling program
Variable<int> nItem = Variable.New<int>().Named("nItem" + s);
// A range over the items
Range item = new Range(nItem).Named("item" + s);
// An array of feature vectors - their observed values will be
// set by the calling program
VariableArray<Vector> xValues = Variable.Array<Vector>(item);
// The weight vectors for each component
Variable<Vector>[] w = new Variable<Vector>[nClass];
// The prior distribution for weight vector for each component. When
// <see cref="Test"/> is called, this is set to the posterior weight
// distributions from <see cref="Train"/>
Variable<VectorGaussian>[] wPrior = new Variable<VectorGaussian>[nClass];
// Loop over the classes
for (int c = 0; c < nClass; c++)
{
// The priors will be set by the calling program
wPrior[c] = Variable.New<VectorGaussian>();
// The weights are sampled from the prior distributions
w[c] = Variable<Vector>.Random(wPrior[c]);
}
// Loop over the data
VariableArray<int> ytest = Variable.Array<int>(item).Named("ytest" + s);
using (Variable.ForEach(item))
{
// The score for this item across all components
Variable<double>[] score = BPMUtils.ComputeClassScores(w, xValues[item], NoisePrec);
// The constraints on the output variable
ytest[item] = Variable.DiscreteUniform(nClass);
BPMUtils.ConstrainMaximum(ytest[item], score, nClass);
}
// Store the variables
BPMVarsForTest bpmVar = new BPMVarsForTest();
bpmVar.ie = new InferenceEngine();
bpmVar.xValues = xValues;
bpmVar.y = ytest;
bpmVar.nItems = nItem;
bpmVar.w = w;
bpmVar.wPrior = wPrior;
return bpmVar;
}