/// <summary>
/// Generates a random vector of observations from the model.
/// </summary>
///
/// <param name="samples">The number of samples to generate.</param>
/// <param name="logLikelihood">The log-likelihood of the generated observation sequence.</param>
/// <param name="path">The Viterbi path of the generated observation sequence.</param>
///
/// <returns>A random vector of observations drawn from the model.</returns>
///
public TObservation[] Generate(int samples, out int[] path, out double logLikelihood)
{
double[] transitions = LogInitial;
logLikelihood = 0; // log(1)
path = new int[samples];
TObservation[] observations = new TObservation[samples];
// For each observation to be generated
for (int t = 0; t < observations.Length; t++)
{
// Navigate randomly on one of the state transitions
int state = GeneralDiscreteDistribution.Random(Elementwise.Exp(transitions));
// Generate a sample for the state
var sampling = (ISampleableDistribution <TObservation>)Emissions[state];
TObservation symbol = sampling.Generate();
// Store the sample
observations[t] = symbol;
path[t] = state;
// Compute log-likelihood up to this point
logLikelihood += transitions[state] +
Emissions[state].LogProbabilityFunction(symbol);
// Continue sampling
transitions = LogTransitions.GetRow(state);
}
return(observations);
}
/// <summary>
/// Creates a new object that is a copy of the current instance.
/// </summary>
///
/// <returns>
/// A new object that is a copy of this instance.
/// </returns>
///
public override object Clone()
{
double[][] A = LogTransitions.Copy();
double[][] B = LogEmissions.Copy();
double[] pi = LogInitial.Copy();
return(new HiddenMarkovModel(A, B, pi, logarithm: true));
}
/// <summary>
/// Creates a new object that is a copy of the current instance.
/// </summary>
///
/// <returns>
/// A new object that is a copy of this instance.
/// </returns>
///
public virtual object Clone()
{
double[][] A = LogTransitions.Copy();
double[] pi = LogInitial.Copy();
TDistribution[] B = Emissions.MemberwiseClone();
return(new HiddenMarkovModel <TDistribution, TObservation>(A, B, pi, logarithm: true));
}
/// <summary>
/// Creates a new object that is a copy of the current instance.
/// </summary>
///
/// <returns>
/// A new object that is a copy of this instance.
/// </returns>
///
public override object Clone()
{
double[,] A = (double[, ])LogTransitions.Clone();
double[,] B = (double[, ])Emissions.Clone();
double[] pi = (double[])LogInitial.Clone();
return(new HiddenMarkovModel(A, B, pi, logarithm: true));
}
public HiddenMarkovModel <GeneralDiscreteDistribution> ToContinuousModel()
{
var transitions = (double[, ])LogTransitions.Clone();
var probabilities = (double[])LogInitial.Clone();
var emissions = new GeneralDiscreteDistribution[States];
for (int i = 0; i < emissions.Length; i++)
{
emissions[i] = new GeneralDiscreteDistribution(Accord.Math.Matrix.GetRow(Emissions, i));
}
return(new HiddenMarkovModel <GeneralDiscreteDistribution>(transitions, emissions, probabilities));
}