de.jstacs.scoringFunctions.homogeneous
Class HMMScoringFunction

java.lang.Object
  de.jstacs.scoringFunctions.AbstractNormalizableScoringFunction
      de.jstacs.scoringFunctions.AbstractVariableLengthScoringFunction
          de.jstacs.scoringFunctions.homogeneous.HomogeneousScoringFunction
              de.jstacs.scoringFunctions.homogeneous.HMMScoringFunction

All Implemented Interfaces:

NormalizableScoringFunction, ScoringFunction, VariableLengthScoringFunction, Storable, Cloneable

public class HMMScoringFunction
extends HomogeneousScoringFunction

This scoring function implements a homogeneous Markov model of arbitrary order for any sequence length. The scoring function uses the parameterization of Meila if one uses the free parameters, which yields in a non-concave log conditional likelihood.

Author:

Jens Keilwagen

Field Summary

Fields inherited from class de.jstacs.scoringFunctions.AbstractNormalizableScoringFunction

alphabets, length, r

Fields inherited from interface de.jstacs.scoringFunctions.ScoringFunction

UNKNOWN

Constructor Summary

HMMScoringFunction(AlphabetContainer alphabets, int order, double classEss, double[] sumOfHyperParams, boolean plugIn, boolean optimize, int starts)
This is the main constructor that creates an instance of a homogeneous Markov model of arbitrary order.

HMMScoringFunction(AlphabetContainer alphabets, int order, double classEss, int length)
This is a convenience constructor for creating an instance of a homogeneous Markov model of arbitrary order.

HMMScoringFunction(StringBuffer xml)
This is the constructor for Storable.

Method Summary

void	addGradientOfLogPriorTerm(double[] grad, int start) This method computes the gradient of NormalizableScoringFunction.getLogPriorTerm() for each parameter of this model.
HMMScoringFunction	clone() Creates a clone (deep copy) of the current ScoringFunction instance.
Sample	emit(int numberOfSequences, int... seqLength) This method returns a Sample object containing artificial sequence(s).
protected void	fromXML(StringBuffer xml) This method is called in the constructor for the Storable interface to create a scoring function from a StringBuffer.
double[][][]	getAllConditionalStationaryDistributions() This method returns the stationary conditional distributions.
double[]	getCurrentParameterValues() Returns a double array of dimension ScoringFunction.getNumberOfParameters() containing the current parameter values.
double	getEss() Returns the equivalent sample size (ess) of this model, i.e. the equivalent sample size for the class or component that is represented by this model.
String	getInstanceName() Returns a short instance name.
double	getLogNormalizationConstant(int length) This method returns the logarithm of the normalization constant for a given sequence length.
double	getLogPartialNormalizationConstant(int parameterIndex, int length) This method returns the logarithm of the partial normalization constant for a given parameter index and a sequence length.
double	getLogPriorTerm() This method computes a value that is proportional to NormalizableScoringFunction.getEss() * NormalizableScoringFunction.getLogNormalizationConstant() + Math.log( prior ) where prior is the prior for the parameters of this model.
double	getLogScore(Sequence seq, int start, int length) This method computes the logarithm of the score for a given subsequence.
double	getLogScoreAndPartialDerivation(Sequence seq, int start, int length, IntList indices, DoubleList dList) This method computes the logarithm of the score and the partial derivations for a given subsequence.
int	getMaximalMarkovOrder() Returns the maximal used markov oder.
int	getNumberOfParameters() Returns the number of parameters in this ScoringFunction.
int	getNumberOfRecommendedStarts() This method returns the number of recommended optimization starts.
int	getSizeOfEventSpaceForRandomVariablesOfParameter(int index) Returns the size of the event space of the random variables that are affected by parameter no.
static double[]	getSumOfHyperParameters(int order, int length, double ess) This method returns an array that can be used in the constructor HMMScoringFunction(AlphabetContainer, int, double, double[], boolean, boolean, int) containing the sums of the specific hyperparameters.
void	initializeFunction(int index, boolean freeParams, Sample[] data, double[][] weights) This method creates the underlying structure of the ScoringFunction.
void	initializeFunctionRandomly(boolean freeParams) This method initializes the ScoringFunction randomly.
void	initializeUniformly(boolean freeParams) This method allows to initialize the instance with an uniform distribution.
boolean	isInitialized() This method can be used to determine whether the model is initialized.
boolean	isNormalized() This method indicates whether the implemented score is already normalized to 1 or not.
void	setParameterOptimization(boolean optimize) This method allows the user to specify whether the parameters should be optimized or not.
void	setParameters(double[] params, int start) This method sets the internal parameters to the values of params between start and start + ScoringFunction.getNumberOfParameters() - 1
void	setStatisticForHyperparameters(int[] length, double[] weight) This method sets the hyperparameters for the model parameters by evaluating the given statistic.
String	toString()
StringBuffer	toXML() This method returns an XML representation as StringBuffer of an instance of the implementing class.

Methods inherited from class de.jstacs.scoringFunctions.AbstractVariableLengthScoringFunction

getLogNormalizationConstant, getLogPartialNormalizationConstant, getLogScore, getLogScoreAndPartialDerivation

Methods inherited from class de.jstacs.scoringFunctions.AbstractNormalizableScoringFunction

getAlphabetContainer, getInitialClassParam, getLength, getLogScore, getLogScoreAndPartialDerivation, getNumberOfStarts, isNormalized

Methods inherited from class java.lang.Object

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

Methods inherited from interface de.jstacs.scoringFunctions.NormalizableScoringFunction

getInitialClassParam

Methods inherited from interface de.jstacs.scoringFunctions.ScoringFunction

getAlphabetContainer, getLength, getLogScore, getLogScoreAndPartialDerivation

Constructor Detail

HMMScoringFunction

public HMMScoringFunction(AlphabetContainer alphabets,
                          int order,
                          double classEss,
                          int length)

This is a convenience constructor for creating an instance of a homogeneous Markov model of arbitrary order.

Parameters:

alphabets - the AlphabetContainer

order - the oder of the model (has to be non-negative)

classEss - the equivalent sample size (ess) of the class

length - the sequence length (only used for computing the hyperparameters)

See Also:

getSumOfHyperParameters(int, int, double), HMMScoringFunction(AlphabetContainer, int, double, double[], boolean, boolean, int)

HMMScoringFunction

public HMMScoringFunction(AlphabetContainer alphabets,
                          int order,
                          double classEss,
                          double[] sumOfHyperParams,
                          boolean plugIn,
                          boolean optimize,
                          int starts)

This is the main constructor that creates an instance of a homogeneous Markov model of arbitrary order.

Parameters:

alphabets - the AlphabetContainer

order - the oder of the model (has to be non-negative)

classEss - the equivalent sample size (ess) of the class

sumOfHyperParams - the sum of the hyperparameters for each order (length has to be order, each entry has to be non-negative)

plugIn - a switch which enables to use the MAP-parameters as plug-in parameters

optimize - a switch which enables to optimize or fix the parameters

starts - the number of recommended starts

HMMScoringFunction

public HMMScoringFunction(StringBuffer xml)
                   throws NonParsableException

This is the constructor for Storable. Creates a new HMMScoringFunction out of its XML representation as returned by fromXML(StringBuffer).

Parameters:

xml - the XML representation as StringBuffer

Throws:

NonParsableException - if the StringBuffer representation could not be parsed

Method Detail

getSumOfHyperParameters

public static double[] getSumOfHyperParameters(int order,
                                               int length,
                                               double ess)

This method returns an array that can be used in the constructor HMMScoringFunction(AlphabetContainer, int, double, double[], boolean, boolean, int) containing the sums of the specific hyperparameters.

Parameters:

order - the order of the model

length - the sequence length

ess - the class ESS

Returns:

an array containing the sums of the specific hyperparameters

See Also:

HMMScoringFunction(AlphabetContainer, int, double, double[], boolean, boolean, int)

clone

public HMMScoringFunction clone()
                         throws CloneNotSupportedException

Description copied from interface: ScoringFunction

Creates a clone (deep copy) of the current ScoringFunction instance.

Specified by:

clone in interface ScoringFunction

Overrides:

clone in class AbstractNormalizableScoringFunction

Returns:

the cloned instance of the current ScoringFunction

Throws:

CloneNotSupportedException - if something went wrong while cloning the ScoringFunction

getInstanceName

public String getInstanceName()

Description copied from interface: ScoringFunction

Returns a short instance name.

Returns:

a short instance name

getLogScore

public double getLogScore(Sequence seq,
                          int start,
                          int length)

Description copied from interface: VariableLengthScoringFunction

This method computes the logarithm of the score for a given subsequence.

Parameters:

seq - the Sequence

start - the start index in the Sequence

length - the length of the Sequence beginning at start

Returns:

the logarithm of the score for the subsequence

See Also:

ScoringFunction.getLogScore(Sequence, int)

getLogScoreAndPartialDerivation

public double getLogScoreAndPartialDerivation(Sequence seq,
                                              int start,
                                              int length,
                                              IntList indices,
                                              DoubleList dList)

Description copied from interface: VariableLengthScoringFunction

This method computes the logarithm of the score and the partial derivations for a given subsequence.

Parameters:

seq - the Sequence

start - the start index in the Sequence

length - the end index in the Sequence

indices - an IntList of indices, after method invocation the list should contain the indices i where is not zero

dList - a DoubleList of partial derivations, after method invocation the list should contain the corresponding that are not zero

Returns:

the logarithm of the score

See Also:

ScoringFunction.getLogScoreAndPartialDerivation(Sequence, int, IntList, DoubleList)

getNumberOfParameters

public int getNumberOfParameters()

Description copied from interface: ScoringFunction

Returns the number of parameters in this ScoringFunction. If the number of parameters is not known yet, the method returns ScoringFunction.UNKNOWN.

Returns:

the number of parameters in this ScoringFunction

See Also:

ScoringFunction.UNKNOWN

setParameters

public void setParameters(double[] params,
                          int start)

Description copied from interface: ScoringFunction

This method sets the internal parameters to the values of params between start and start + ScoringFunction.getNumberOfParameters() - 1

Parameters:

params - the new parameters

start - the start index in params

toXML

public StringBuffer toXML()

Description copied from interface: Storable

This method returns an XML representation as StringBuffer of an instance of the implementing class.

Returns:

the XML representation

getCurrentParameterValues

public double[] getCurrentParameterValues()

Description copied from interface: ScoringFunction

Returns a double array of dimension ScoringFunction.getNumberOfParameters() containing the current parameter values. If one likes to use these parameters to start an optimization it is highly recommended to invoke ScoringFunction.initializeFunction(int, boolean, Sample[], double[][]) before. After an optimization this method can be used to get the current parameter values.

Returns:

the current parameter values

initializeFunction

public void initializeFunction(int index,
                               boolean freeParams,
                               Sample[] data,
                               double[][] weights)

Description copied from interface: ScoringFunction

This method creates the underlying structure of the ScoringFunction.

Parameters:

index - the index of the class the ScoringFunction models

freeParams - indicates whether the (reduced) parameterization is used

data - the samples

weights - the weights of the sequences in the samples

initializeFunctionRandomly

public void initializeFunctionRandomly(boolean freeParams)

Description copied from interface: ScoringFunction

This method initializes the ScoringFunction randomly. It has to create the underlying structure of the ScoringFunction.

Parameters:

freeParams - indicates whether the (reduced) parameterization is used

fromXML

protected void fromXML(StringBuffer xml)
                throws NonParsableException

Description copied from class: AbstractNormalizableScoringFunction

This method is called in the constructor for the Storable interface to create a scoring function from a StringBuffer.

Specified by:

fromXML in class AbstractNormalizableScoringFunction

Parameters:

xml - the XML representation as StringBuffer

Throws:

NonParsableException - if the StringBuffer could not be parsed

See Also:

AbstractNormalizableScoringFunction.AbstractNormalizableScoringFunction(StringBuffer)

getSizeOfEventSpaceForRandomVariablesOfParameter

public int getSizeOfEventSpaceForRandomVariablesOfParameter(int index)

Description copied from interface: NormalizableScoringFunction

Returns the size of the event space of the random variables that are affected by parameter no. index, i.e. the product of the sizes of the alphabets at the position of each random variable affected by parameter index. For DNA alphabets this corresponds to 4 for a PWM, 16 for a WAM except position 0, ...

Parameters:

index - the index of the parameter

Returns:

the size of the event space

getLogNormalizationConstant

public double getLogNormalizationConstant(int length)

Description copied from interface: VariableLengthScoringFunction

This method returns the logarithm of the normalization constant for a given sequence length.

Parameters:

length - the sequence length

Returns:

the logarithm of the normalization constant

See Also:

NormalizableScoringFunction.getLogNormalizationConstant()

getLogPartialNormalizationConstant

public double getLogPartialNormalizationConstant(int parameterIndex,
                                                 int length)
                                          throws Exception

Description copied from interface: VariableLengthScoringFunction

This method returns the logarithm of the partial normalization constant for a given parameter index and a sequence length.

Parameters:

parameterIndex - the index of the parameter

length - the sequence length

Returns:

the logarithm of the partial normalization constant

Throws:

Exception - if something went wrong

See Also:

NormalizableScoringFunction.getLogPartialNormalizationConstant(int)

getEss

public double getEss()

Description copied from interface: NormalizableScoringFunction

Returns the equivalent sample size (ess) of this model, i.e. the equivalent sample size for the class or component that is represented by this model.

Returns:

the equivalent sample size.

toString

public String toString()

Overrides:

toString in class Object

getLogPriorTerm

public double getLogPriorTerm()

Description copied from interface: NormalizableScoringFunction

This method computes a value that is proportional to

NormalizableScoringFunction.getEss() * NormalizableScoringFunction.getLogNormalizationConstant() + Math.log( prior )

where prior is the prior for the parameters of this model.

Returns:

a value that is proportional to NormalizableScoringFunction.getEss() * NormalizableScoringFunction.getLogNormalizationConstant() + Math.log( prior ).

See Also:

NormalizableScoringFunction.getEss(), NormalizableScoringFunction.getLogNormalizationConstant()

addGradientOfLogPriorTerm

public void addGradientOfLogPriorTerm(double[] grad,
                                      int start)

Description copied from interface: NormalizableScoringFunction

This method computes the gradient of NormalizableScoringFunction.getLogPriorTerm() for each parameter of this model. The results are added to the array grad beginning at index start.

Parameters:

grad - the array of gradients

start - the start index in the grad array, where the partial derivations for the parameters of this models shall be entered

See Also:

NormalizableScoringFunction.getLogPriorTerm()

isNormalized

public boolean isNormalized()

Description copied from interface: NormalizableScoringFunction

This method indicates whether the implemented score is already normalized to 1 or not. The standard implementation returns false.

Specified by:

isNormalized in interface NormalizableScoringFunction

Overrides:

isNormalized in class AbstractNormalizableScoringFunction

Returns:

true if the implemented score is already normalized to 1, false otherwise

isInitialized

public boolean isInitialized()

Description copied from interface: ScoringFunction

This method can be used to determine whether the model is initialized. If the model is not initialized you should invoke the method ScoringFunction.initializeFunction(int, boolean, Sample[], double[][]).

Returns:

true if the model is initialized, false otherwise

getMaximalMarkovOrder

public int getMaximalMarkovOrder()

Description copied from class: HomogeneousScoringFunction

Returns the maximal used markov oder.

Specified by:

getMaximalMarkovOrder in class HomogeneousScoringFunction

Returns:

the maximal used markov oder

getNumberOfRecommendedStarts

public int getNumberOfRecommendedStarts()

Description copied from interface: ScoringFunction

This method returns the number of recommended optimization starts. The standard implementation returns 1.

Specified by:

getNumberOfRecommendedStarts in interface ScoringFunction

Overrides:

getNumberOfRecommendedStarts in class AbstractNormalizableScoringFunction

Returns:

the number of recommended optimization starts

setParameterOptimization

public void setParameterOptimization(boolean optimize)

This method allows the user to specify whether the parameters should be optimized or not.

Parameters:

optimize - indicates if the parameters should be optimized or not

getAllConditionalStationaryDistributions

public double[][][] getAllConditionalStationaryDistributions()

This method returns the stationary conditional distributions. The first dimension of the result is used for the order, the second is used for encoding the context, and the third is used for the different values of the random variable.

For an homogeneous Markov model of order 2 it returns an array containing the stationary symbol distribution as first entry, the conditional stationary distribution of order 1 as second entry and the conditional distribution of order 2 as third entry.

Returns:

all conditional stationary distributions

setStatisticForHyperparameters

public void setStatisticForHyperparameters(int[] length,
                                           double[] weight)
                                    throws Exception

Description copied from interface: VariableLengthScoringFunction

This method sets the hyperparameters for the model parameters by evaluating the given statistic. The statistic can be interpreted as follows: The model has seen a number of sequences. From these sequences it is only known how long (length) and how often ( weight) they have been seen.

Parameters:

length - the non-negative lengths of the sequences

weight - the non-negative weight for the corresponding sequence

Throws:

Exception - if something went wrong

See Also:

Mutable

emit

public Sample emit(int numberOfSequences,
                   int... seqLength)
            throws Exception

This method returns a Sample object containing artificial sequence(s).

There are 2 different possibilities to create a Sample:

1. emitSample( int n, int l ) returns a Sample with n sequences of length l.
2. emitSample( int n, int[] l ) should return a Sample with n sequences which have a sequence length corresponding to the entry in the array.

Parameters:

numberOfSequences - the number of sequences that should be contained in the returned Sample

seqLength - the length of the sequences

Returns:

a Sample containing numberOfSequences artificial sequence(s)

Throws:

Exception - if the emission of the artificial Sample did not succeed

See Also:

Sample

initializeUniformly

public void initializeUniformly(boolean freeParams)

Description copied from class: HomogeneousScoringFunction

This method allows to initialize the instance with an uniform distribution.

Specified by:

initializeUniformly in class HomogeneousScoringFunction

Parameters:

freeParams - a switch whether to take only free parameters or to take all