org.encog.ml.hmm.distributions
Interface StateDistribution

All Superinterfaces:
Cloneable, Serializable
All Known Implementing Classes:
ContinousDistribution, DiscreteDistribution

public interface StateDistribution
extends Cloneable, Serializable

This class represents a "state distribution". This is the means by which the probabilities between the states and observations are mapped. Currently two are supported. Use ContinousDistribution to use a Gaussian-based continuous distribution. Use DiscreteDistribution for a item-based distribution.


Method Summary
 StateDistribution clone()
           
 void fit(MLDataSet set)
          Fit this distribution to the specified data set.
 void fit(MLDataSet set, double[] weights)
          Fit this distribution to the specified data set, given the specified weights, per element.
 MLDataPair generate()
          Generate a random data pair, based on the probabilities.
 double probability(MLDataPair o)
          Determine the probability of the specified data pair.
 

Method Detail

clone

StateDistribution clone()
Returns:
A clone of this distribution.

fit

void fit(MLDataSet set)
Fit this distribution to the specified data set.

Parameters:
set - The data set to fit to.

fit

void fit(MLDataSet set,
         double[] weights)
Fit this distribution to the specified data set, given the specified weights, per element.

Parameters:
set - The data set to fit to.
weights - The weights.

generate

MLDataPair generate()
Generate a random data pair, based on the probabilities.

Returns:
A random data pair.

probability

double probability(MLDataPair o)
Determine the probability of the specified data pair.

Parameters:
o - The pair to consider.
Returns:
The probability.


Copyright © 2014. All Rights Reserved.