de.jstacs.sequenceScores.statisticalModels.trainable.mixture

Class MixtureTrainSM

java.lang.Object

de.jstacs.sequenceScores.statisticalModels.trainable.AbstractTrainableStatisticalModel

de.jstacs.sequenceScores.statisticalModels.trainable.mixture.AbstractMixtureTrainSM

de.jstacs.sequenceScores.statisticalModels.trainable.mixture.MixtureTrainSM

All Implemented Interfaces:

SequenceScore, StatisticalModel, TrainableStatisticalModel, Storable, Cloneable

Direct Known Subclasses:

SharedStructureMixture

public class MixtureTrainSM
extends AbstractMixtureTrainSM

The class for a mixture model of any TrainableStatisticalModels.

If you use Gibbs sampling temporary files will be created in the Java temp folder. These files will be deleted if no reference to the current instance exists and the Garbage Collector is called. Therefore it is recommended to call the Garbage Collector explicitly at the end of any application.

Author:

Jens Keilwagen, Berit Haldemann

Nested Class Summary

Nested classes/interfaces inherited from class de.jstacs.sequenceScores.statisticalModels.trainable.mixture.AbstractMixtureTrainSM

AbstractMixtureTrainSM.Algorithm, AbstractMixtureTrainSM.Parameterization

Field Summary

Fields inherited from class de.jstacs.sequenceScores.statisticalModels.trainable.mixture.AbstractMixtureTrainSM

algorithm, algorithmHasBeenRun, alternativeModel, best, burnInTest, componentHyperParams, compProb, counter, dimension, estimateComponentProbs, file, filereader, filewriter, initialIteration, logWeights, model, optimizeModel, sample, samplingIndex, seqWeights, sostream, starts, stationaryIteration, weights

Fields inherited from class de.jstacs.sequenceScores.statisticalModels.trainable.AbstractTrainableStatisticalModel

alphabets, length

Constructor Summary

Constructors

Modifier	Constructor and Description
	MixtureTrainSM(int length, TrainableStatisticalModel[] models, double[] weights, int starts, double alpha, TerminationCondition tc, AbstractMixtureTrainSM.Parameterization parametrization) Creates an instance using EM and fixed component probabilities.
	MixtureTrainSM(int length, TrainableStatisticalModel[] models, double[] weights, int starts, int initialIteration, int stationaryIteration, BurnInTest burnInTest) Creates an instance using Gibbs Sampling and fixed component probabilities.
protected	MixtureTrainSM(int length, TrainableStatisticalModel[] models, int starts, boolean estimateComponentProbs, double[] componentHyperParams, double[] weights, AbstractMixtureTrainSM.Algorithm algorithm, double alpha, TerminationCondition tc, AbstractMixtureTrainSM.Parameterization parametrization, int initialIteration, int stationaryIteration, BurnInTest burnInTest) Creates a new MixtureTrainSM.
	MixtureTrainSM(int length, TrainableStatisticalModel[] models, int starts, double[] componentHyperParams, double alpha, TerminationCondition tc, AbstractMixtureTrainSM.Parameterization parametrization) Creates an instance using EM and estimating the component probabilities.
	MixtureTrainSM(int length, TrainableStatisticalModel[] models, int starts, double[] componentHyperParams, int initialIteration, int stationaryIteration, BurnInTest burnInTest) Creates an instance using Gibbs Sampling and sampling the component probabilities.
	MixtureTrainSM(StringBuffer xml) The constructor for the interface Storable.

Method Summary

Methods

Modifier and Type	Method and Description
double[][]	doFirstIteration(DataSet data, double[] dataWeights, double[][] partitioning) This method enables you to train a mixture model with a fixed start partitioning.
protected double[][]	doFirstIteration(double[] dataWeights, MultivariateRandomGenerator m, MRGParams[] params) This method will do the first step in the train algorithm for the current model on the internal data set.
protected Sequence[]	emitDataSetUsingCurrentParameterSet(int n, int... lengths) The method returns an array of sequences using the current parameter set.
protected double	getLogProbUsingCurrentParameterSetFor(int component, Sequence s, int start, int end) Returns the logarithmic probability for the sequence and the given component using the current parameter set.
protected double	getNewWeights(double[] dataWeights, double[] w, double[][] seqweights) Computes sequence weights and returns the score.
protected void	setTrainData(DataSet data) This method is invoked by the train-method and sets for a given data set the data set that should be used for train.
String	toString(NumberFormat nf) This method returns a String representation of the instance.

Methods inherited from class de.jstacs.sequenceScores.statisticalModels.trainable.mixture.AbstractMixtureTrainSM

algorithmHasBeenRun, checkLength, checkModelsForGibbsSampling, clone, continueIterations, continueIterations, createSeqWeightsArray, doFirstIteration, doFirstIteration, draw, emitDataSet, extendSampling, extractFurtherInformation, finalize, fromXML, getCharacteristics, getFurtherInformation, getIndexOfMaximalComponentFor, getInstanceName, getLogPriorTerm, getLogPriorTermForComponentProbs, getLogProbFor, getLogProbFor, getLogScoreFor, getModel, getModels, getMRG, getMRGParams, getNameOfAlgorithm, getNewComponentProbs, getNewParameters, getNewParametersForModel, getNumberOfComponents, getNumericalCharacteristics, getScoreForBestRun, getWeights, initModelForSampling, initWithPrior, isInitialized, isInSamplingMode, iterate, iterate, max, modifyWeights, parseNextParameterSet, parseParameterSet, samplingStopped, setAlpha, setOutputStream, setWeights, swap, toXML, train

Methods inherited from class de.jstacs.sequenceScores.statisticalModels.trainable.AbstractTrainableStatisticalModel

check, getAlphabetContainer, getLength, getLogProbFor, getLogProbFor, getLogScoreFor, getLogScoreFor, getLogScoreFor, getLogScoreFor, getMaximalMarkovOrder, toString, train

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, wait, wait, wait

Constructor Detail

MixtureTrainSM

protected MixtureTrainSM(int length,
              TrainableStatisticalModel[] models,
              int starts,
              boolean estimateComponentProbs,
              double[] componentHyperParams,
              double[] weights,
              AbstractMixtureTrainSM.Algorithm algorithm,
              double alpha,
              TerminationCondition tc,
              AbstractMixtureTrainSM.Parameterization parametrization,
              int initialIteration,
              int stationaryIteration,
              BurnInTest burnInTest)
                  throws IllegalArgumentException,
                         WrongAlphabetException,
                         CloneNotSupportedException

Creates a new MixtureTrainSM. This constructor can be used for any algorithm since it takes all necessary values as parameters.

Parameters:

length - the length used in this model

models - the single models building the MixtureTrainSM, if the model is trained using AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING the models that will be adjusted have to implement SamplingComponent

starts - the number of times the algorithm will be started in the train-method, at least 1

estimateComponentProbs - the switch for estimating the component probabilities in the algorithm or to hold them fixed; if the component parameters are fixed, the values of weights will be used, otherwise the componentHyperParams will be incorporated in the adjustment

componentHyperParams - the hyperparameters for the component assignment prior

• will only be used if estimateComponentProbs == true
• the array has to be null or has to have length models.length
• null or an array with all values zero (0) then ML
• otherwise (all values positive) a prior is used (MAP, MP, ...)
• depends on the parameterization

weights - null or the weights for the components (then weights.length == models.length)

algorithm - either AbstractMixtureTrainSM.Algorithm.EM or AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING

alpha - only for AbstractMixtureTrainSM.Algorithm.EM
the positive parameter for the Dirichlet distribution which is used when you invoke train to initialize the gammas. It is recommended to use alpha = 1 (uniform distribution on a simplex).

tc - only for AbstractMixtureTrainSM.Algorithm.EM
the TerminationCondition for stopping the EM-algorithm, tc has to return true from TerminationCondition.isSimple()

parametrization - only for AbstractMixtureTrainSM.Algorithm.EM
the type of the component probability parameterization;

• AbstractMixtureTrainSM.Parameterization.THETA or AbstractMixtureTrainSM.Parameterization.LAMBDA
• the parameterization of a component is determined by the component model
• it is recommended to use the same parameterization for the components and the component assignment probabilities
• it is recommended to use AbstractMixtureTrainSM.Parameterization.LAMBDA

initialIteration - only for AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING
the positive length of the initial sampling phase (at least 1, at most stationaryIteration/starts)

stationaryIteration - only for AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING
the positive length of the stationary phase (at least 1) (summed over all starts), i.e. the number of parameter sets that is used for approximation

burnInTest - only for AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING
the test that will be used to determine the length of the burn-in phase

Throws:

IllegalArgumentException - if

• the models are not able to score the sequence of length length
• dimension < 1
• weights != null && weights.length != dimension
• weights != null and it exists an i where weights[i] < 0
• starts < 1
• componentHyperParams are not correct
• the algorithm specific parameters are not correct

WrongAlphabetException - if not all models work on the same alphabet

CloneNotSupportedException - if the models can not be cloned

MixtureTrainSM

public MixtureTrainSM(int length,
              TrainableStatisticalModel[] models,
              int starts,
              double[] componentHyperParams,
              double alpha,
              TerminationCondition tc,
              AbstractMixtureTrainSM.Parameterization parametrization)
               throws IllegalArgumentException,
                      WrongAlphabetException,
                      CloneNotSupportedException

Creates an instance using EM and estimating the component probabilities.

Parameters:

length - the length used in this model

models - the single models building the MixtureTrainSM, if the model is trained using AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING the models that will be adjusted have to implement SamplingComponent

starts - the number of times the algorithm will be started in the train-method, at least 1

componentHyperParams - the hyperparameters for the component assignment prior

• will only be used if estimateComponentProbs == true
• the array has to be null or has to have length models.length
• null or an array with all values zero (0) then ML
• otherwise (all values positive) a prior is used (MAP, MP, ...)
• depends on the parameterization

alpha - only for AbstractMixtureTrainSM.Algorithm.EM
the positive parameter for the Dirichlet distribution which is used when you invoke train to initialize the gammas. It is recommended to use alpha = 1 (uniform distribution on a simplex).

tc - only for AbstractMixtureTrainSM.Algorithm.EM
the TerminationCondition for stopping the EM-algorithm, tc has to return true from TerminationCondition.isSimple()

parametrization - only for AbstractMixtureTrainSM.Algorithm.EM
the type of the component probability parameterization

• AbstractMixtureTrainSM.Parameterization.THETA or AbstractMixtureTrainSM.Parameterization.LAMBDA
• the parameterization of a component is determined by the component model
• it is recommended to use the same parameterization for the components and the component assignment probabilities
• it is recommended to use AbstractMixtureTrainSM.Parameterization.LAMBDA

Throws:

IllegalArgumentException - if

• the models are not able to score the sequence of length length
• dimension < 1
• weights != null && weights.length != dimension
• weights != null and it exists an i where weights[i] < 0
• starts < 1
• componentHyperParams are not correct
• the algorithm specific parameters are not correct

WrongAlphabetException - if not all models work on the same alphabet

CloneNotSupportedException - if the models can not be cloned

See Also:

MixtureTrainSM(int, de.jstacs.sequenceScores.statisticalModels.trainable.TrainableStatisticalModel[], int, boolean, double[], double[], de.jstacs.sequenceScores.statisticalModels.trainable.mixture.AbstractMixtureTrainSM.Algorithm, double, TerminationCondition, de.jstacs.sequenceScores.statisticalModels.trainable.mixture.AbstractMixtureTrainSM.Parameterization, int, int, de.jstacs.sampling.BurnInTest), AbstractMixtureTrainSM.Algorithm.EM

MixtureTrainSM

public MixtureTrainSM(int length,
              TrainableStatisticalModel[] models,
              double[] weights,
              int starts,
              double alpha,
              TerminationCondition tc,
              AbstractMixtureTrainSM.Parameterization parametrization)
               throws IllegalArgumentException,
                      WrongAlphabetException,
                      CloneNotSupportedException

Creates an instance using EM and fixed component probabilities.

Parameters:

length - the length used in this model

models - the single models building the MixtureTrainSM, if the model is trained using AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING the models that will be adjusted have to implement SamplingComponent

starts - the number of times the algorithm will be started in the train-method, at least 1

weights - null or the weights for the components (then weights.length == models.length)

alpha - only for AbstractMixtureTrainSM.Algorithm.EM
the positive parameter for the Dirichlet distribution which is used when you invoke train to initialize the gammas. It is recommended to use alpha = 1 (uniform distribution on a simplex).

tc - only for AbstractMixtureTrainSM.Algorithm.EM
the TerminationCondition for stopping the EM-algorithm, tc has to return true from TerminationCondition.isSimple()

parametrization - only for AbstractMixtureTrainSM.Algorithm.EM
the type of the component probability parameterization;

• AbstractMixtureTrainSM.Parameterization.THETA or AbstractMixtureTrainSM.Parameterization.LAMBDA
• the parameterization of a component is determined by the component model
• it is recommended to use the same parameterization for the components and the component assignment probabilities
• it is recommended to use AbstractMixtureTrainSM.Parameterization.LAMBDA

Throws:

IllegalArgumentException - if

• the models are not able to score the sequence of length length
• dimension < 1
• weights != null && weights.length != dimension
• weights != null and it exists an i where weights[i] < 0
• starts < 1
• componentHyperParams are not correct
• the algorithm specific parameters are not correct

WrongAlphabetException - if not all models work on the same alphabet

CloneNotSupportedException - if the models can not be cloned

See Also:

MixtureTrainSM(int, de.jstacs.sequenceScores.statisticalModels.trainable.TrainableStatisticalModel[], int, boolean, double[], double[], de.jstacs.sequenceScores.statisticalModels.trainable.mixture.AbstractMixtureTrainSM.Algorithm, double, TerminationCondition, de.jstacs.sequenceScores.statisticalModels.trainable.mixture.AbstractMixtureTrainSM.Parameterization, int, int, de.jstacs.sampling.BurnInTest), AbstractMixtureTrainSM.Algorithm.EM

MixtureTrainSM

public MixtureTrainSM(int length,
              TrainableStatisticalModel[] models,
              int starts,
              double[] componentHyperParams,
              int initialIteration,
              int stationaryIteration,
              BurnInTest burnInTest)
               throws IllegalArgumentException,
                      WrongAlphabetException,
                      CloneNotSupportedException

Creates an instance using Gibbs Sampling and sampling the component probabilities.

Parameters:

length - the length used in this model

models - the single models building the MixtureTrainSM, if the model is trained using AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING the models that will be adjusted have to implement SamplingComponent

starts - the number of times the algorithm will be started in the train-method, at least 1

componentHyperParams - the hyperparameters for the component assignment prior

• will only be used if estimateComponentProbs == true
• the array has to be null or has to have length models.length
• null or an array with all values zero (0) then ML
• otherwise (all values positive) a prior is used (MAP, MP, ...)
• depends on the parameterization

initialIteration - only for AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING
the positive length of the initial sampling phase (at least 1, at most stationaryIteration/starts)

stationaryIteration - only for AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING
the positive length of the stationary phase (at least 1) (summed over all starts), i.e. the number of parameter sets that is used for approximation

burnInTest - only for AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING
the test that will be used to determine the length of the burn-in phase

Throws:

IllegalArgumentException - if

• the models are not able to score the sequence of length length
• dimension < 1
• weights != null && weights.length != dimension
• weights != null and it exists an i where weights[i] < 0
• starts < 1
• componentHyperParams are not correct
• the algorithm specific parameters are not correct

WrongAlphabetException - if not all models work on the same alphabet

CloneNotSupportedException - if the models can not be cloned

See Also:

MixtureTrainSM(int, de.jstacs.sequenceScores.statisticalModels.trainable.TrainableStatisticalModel[], int, boolean, double[], double[], de.jstacs.sequenceScores.statisticalModels.trainable.mixture.AbstractMixtureTrainSM.Algorithm, double, TerminationCondition, de.jstacs.sequenceScores.statisticalModels.trainable.mixture.AbstractMixtureTrainSM.Parameterization, int, int, de.jstacs.sampling.BurnInTest), AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING

MixtureTrainSM

public MixtureTrainSM(int length,
              TrainableStatisticalModel[] models,
              double[] weights,
              int starts,
              int initialIteration,
              int stationaryIteration,
              BurnInTest burnInTest)
               throws IllegalArgumentException,
                      WrongAlphabetException,
                      CloneNotSupportedException

Creates an instance using Gibbs Sampling and fixed component probabilities.

Parameters:

length - the length used in this model

models - the single models building the MixtureTrainSM, if the model is trained using AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING the models that will be adjusted have to implement SamplingComponent

starts - the number of times the algorithm will be started in the train-method, at least 1

weights - null or the weights for the components (than weights.length == models.length)

initialIteration - only for AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING
the positive length of the initial sampling phase (at least 1, at most stationaryIteration/starts)

stationaryIteration - only for AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING
the positive length of the stationary phase (at least 1) (summed over all starts), i.e. the number of parameter sets that is used for approximation

burnInTest - only for AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING
the test that will be used to determine the length of the burn-in phase

Throws:

IllegalArgumentException - if

• the models are not able to score the sequence of length length
• dimension < 1
• weights != null && weights.length != dimension
• weights != null and it exists an i where weights[i] < 0
• starts < 1
• componentHyperParams are not correct
• the algorithm specific parameters are not correct

WrongAlphabetException - if not all models work on the same alphabet

CloneNotSupportedException - if the models can not be cloned

See Also:

MixtureTrainSM(int, de.jstacs.sequenceScores.statisticalModels.trainable.TrainableStatisticalModel[], int, boolean, double[], double[], de.jstacs.sequenceScores.statisticalModels.trainable.mixture.AbstractMixtureTrainSM.Algorithm, double, TerminationCondition, de.jstacs.sequenceScores.statisticalModels.trainable.mixture.AbstractMixtureTrainSM.Parameterization, int, int, de.jstacs.sampling.BurnInTest), AbstractMixtureTrainSM.Algorithm.GIBBS_SAMPLING

MixtureTrainSM

public MixtureTrainSM(StringBuffer xml)
               throws NonParsableException

The constructor for the interface Storable. Creates a new MixtureTrainSM out of its XML representation.

Parameters:

xml - the XML representation of the model as StringBuffer

Throws:

NonParsableException - if the StringBuffer is not parsable

Method Detail

emitDataSetUsingCurrentParameterSet

protected Sequence[] emitDataSetUsingCurrentParameterSet(int n,
                                             int... lengths)
                                                  throws Exception

Description copied from class: AbstractMixtureTrainSM

The method returns an array of sequences using the current parameter set.

Specified by:

emitDataSetUsingCurrentParameterSet in class AbstractMixtureTrainSM

Parameters:

n - the number of sequences to be sampled

lengths - the corresponding lengths

Returns:

an array of sequences

Throws:

Exception - if it was impossible to sample the sequences

See Also:

StatisticalModel.emitDataSet(int, int...)

doFirstIteration

protected double[][] doFirstIteration(double[] dataWeights,
                          MultivariateRandomGenerator m,
                          MRGParams[] params)
                               throws Exception

Description copied from class: AbstractMixtureTrainSM

This method will do the first step in the train algorithm for the current model on the internal data set. The initialization will be done by randomly setting the component membership. This is useful when nothing is known about the problem.

Specified by:

doFirstIteration in class AbstractMixtureTrainSM

Parameters:

dataWeights - null or the weights of each element of the data set

m - the multivariate random generator

params - the parameters for the multivariate random generator

Returns:

the weighting array used to initialize, this array can be reused in the following iterations

Throws:

Exception - if something went wrong

doFirstIteration

public double[][] doFirstIteration(DataSet data,
                          double[] dataWeights,
                          double[][] partitioning)
                            throws Exception

This method enables you to train a mixture model with a fixed start partitioning. This is useful to compare implementations or if one has a hypothesis how the components should look like.

Parameters:

data - the data set of sequences

dataWeights - null or the weights of each element of the data set

partitioning - a kind of partitioning

1. partitioning.length has to be data.getNumberofElements()
2. for all i: partitioning[i].length has to be getNumberOfModels()
3. 

Returns:

the weighting array used to initialize, this array can be reused in the following iterations

Throws:

Exception - if something went wrong or if the number of components is 1

getLogProbUsingCurrentParameterSetFor

protected double getLogProbUsingCurrentParameterSetFor(int component,
                                           Sequence s,
                                           int start,
                                           int end)
                                                throws Exception

Description copied from class: AbstractMixtureTrainSM

Returns the logarithmic probability for the sequence and the given component using the current parameter set.

Specified by:

getLogProbUsingCurrentParameterSetFor in class AbstractMixtureTrainSM

Parameters:

component - the index of the component

s - the sequence

start - the start position in the sequence

end - the end position in the sequence

Returns:

log P(s,component) = log P(s|component) + log P(component)

Throws:

Exception - if not trained yet or something else went wrong

See Also:

AbstractMixtureTrainSM.getNumberOfComponents()

toString

public String toString(NumberFormat nf)

Description copied from interface: SequenceScore

This method returns a String representation of the instance.

Parameters:

nf - the NumberFormat for the String representation of parameters or probabilities

Returns:

a String representation of the instance

getNewWeights

protected double getNewWeights(double[] dataWeights,
                   double[] w,
                   double[][] seqweights)
                        throws Exception

Computes sequence weights and returns the score.

Specified by:

getNewWeights in class AbstractMixtureTrainSM

Parameters:

dataWeights - the weights for the internal data set (should not be changed)

w - the array for the statistic of the component parameters (shall be filled)

seqweights - an array containing for each component the weights for each sequence (shall be filled)

Returns:

the score

Throws:

Exception - if something went wrong

setTrainData

protected void setTrainData(DataSet data)

Description copied from class: AbstractMixtureTrainSM

This method is invoked by the train-method and sets for a given data set the data set that should be used for train.

Specified by:

setTrainData in class AbstractMixtureTrainSM

Parameters:

data - the given data set of sequences