Package de.jstacs.sequenceScores.statisticalModels.differentiable

Provides all DifferentiableStatisticalModels, which can compute the gradient with respect to their parameters for a given input Sequence.

See:
Description

Interface Summary

DifferentiableStatisticalModel	The interface for normalizable DifferentiableSequenceScores.
SamplingDifferentiableStatisticalModel	Interface for DifferentiableStatisticalModels that can be used for Metropolis-Hastings sampling in a SamplingScoreBasedClassifier.
VariableLengthDiffSM	This is an interface for all DifferentiableStatisticalModels that allow to score subsequences of arbitrary length.

Class Summary

AbstractDifferentiableStatisticalModel	This class is the main part of any ScoreClassifier.
AbstractVariableLengthDiffSM	This abstract class implements some methods declared in DifferentiableStatisticalModel based on the declaration of methods in VariableLengthDiffSM.
CyclicMarkovModelDiffSM	This scoring function implements a cyclic Markov model of arbitrary order and periodicity for any sequence length.
DifferentiableStatisticalModelFactory	This class allows to easily create some frequently used models.
IndependentProductDiffSM	This class enables the user to model parts of a sequence independent of each other.
MappingDiffSM	This class implements a DifferentiableStatisticalModel that works on mapped Sequences.
MarkovRandomFieldDiffSM	This class implements the scoring function for any MRF (Markov Random Field).
MultiDimensionalSequenceWrapperDiffSM	This class implements a simple wrapper for multidimensional sequences.
NormalizedDiffSM	This class makes an unnormalized DifferentiableStatisticalModel to a normalized DifferentiableStatisticalModel.
UniformDiffSM	This DifferentiableStatisticalModel does nothing.

Package de.jstacs.sequenceScores.statisticalModels.differentiable Description

Provides all DifferentiableStatisticalModels, which can compute the gradient with respect to their parameters for a given input Sequence. The parameters of DifferentiableStatisticalModel are learned numerically, typically by gradient-based method like provided in Optimizer.
This is especially used in Jstacs for learning the parameters by discriminative learning principles like maximum conditional likelihood or maximum supervised posterior (see MSPClassifier) or by a unified learning principle (see GenDisMixClassifier).
The sub-package de.jstacs.sequenceScores.statisticalModels.differentiable.directedGraphicalModels contains Bayesian networks and inhomogeneous Markov models.
The sub-package de.jstacs.sequenceScores.statisticalModels.differentiable.homogeneous provides homogeneous models like homogeneous Markov models.
The sub-package de.jstacs.sequenceScores.statisticalModels.differentiable.mixture provides mixture models including an extended ZOOPS model for de-novo motif discovery.
Some of the provided DifferentiableStatisticalModels also implement the interface SamplingDifferentiableStatisticalModel and can be used for Metropolis-Hastings parameter sampling in a SamplingGenDisMixClassifier.