Package be.ac.ulg.montefiore.run.jahmm

This package is an Hidden Markov Model implementation.

See:
          Description

Interface Summary
Centroid<O> The centroid (basically, the mean) of a cluster.
CentroidFactory<O> Creates a centroid for type .
Opdf<O extends Observation> Objects implementing this interface represent an observation probability (distribution) function.
OpdfFactory<D extends Opdf<?>> Classes implementing this interface are able to generate observation probability distribution functions.
 

Class Summary
CentroidObservationInteger This class represents the centroid of a set of ObservationInteger.
CentroidObservationReal This class represents the centroid of a set of ObservationReal.
CentroidObservationVector This class represents the centroid of a set of ObservationVector.
ForwardBackwardCalculator This class can be used to compute the probability of a given observations sequence for a given HMM.
ForwardBackwardScaledCalculator This class can be used to compute the probability of a given observations sequence for a given HMM.
Hmm<O extends Observation> Main Hmm class; it implements an Hidden Markov Model.
KMeansCalculator<K extends CentroidFactory<? super K>> This class can be used to divide a set of elements in clusters using the k-means algorithm.
Observation Observations generated by a Markovian process.
ObservationDiscrete<E extends Enum<E>> This class implements observations whose values are taken out of a finite set implemented as an enumeration.
ObservationInteger This class holds an integer observation.
ObservationReal This class implements observations made of a real value.
ObservationVector This class holds an Observation described by a vector of reals.
OpdfDiscrete<E extends Enum<E>> This class implements a distribution over a finite set of elements.
OpdfDiscreteFactory<E extends Enum<E>> This class can build OpdfInteger observation probability distribution functions.
OpdfGaussian This class represents a (monovariate) gaussian distribution function.
OpdfGaussianFactory This class can build OpdfMultiGaussian observation probability functions.
OpdfGaussianMixture This class implements a mixture of monovariate gaussian distributions.
OpdfGaussianMixtureFactory Implements a factory of Gaussian mixtures distributions.
OpdfInteger This class represents a distribution of a finite number of positive integer observations.
OpdfIntegerFactory This class can build OpdfInteger observation probability functions.
OpdfMultiGaussian This class represents a multivariate gaussian distribution function.
OpdfMultiGaussianFactory This class can build OpdfMultiGaussian observation probability functions.
ViterbiCalculator This class can be used to compute the most probable state sequence matching a given observation sequence (given an HMM).
 

Enum Summary
ForwardBackwardCalculator.Computation Flags used to explain how the observation sequence probability should be computed (either forward, using the alpha array, or backward, using the beta array).
 

Package be.ac.ulg.montefiore.run.jahmm Description

This package is an Hidden Markov Model implementation. It implements the classical related algorithms. The notations are those found in the papers from Rabiner and Juang, namely An introduction to Hidden Markov Model IEEE ASSP (June, 1986) and The segmental k-means algorithm for learn parameters of Hidden Markov Models IEEE ASSP (September, 1990).



Copyright © 2004,2005 Jean-Marc François.