edu.umass.cs.mallet.base.fst

Class HMM

java.lang.Object

edu.umass.cs.mallet.base.fst.Transducer

edu.umass.cs.mallet.base.fst.HMM

All Implemented Interfaces:

Serializable

public class HMM
extends Transducer
implements Serializable

Hidden Markov Model

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	HMM.State
protected static class	HMM.TransitionIterator

Nested classes/interfaces inherited from class edu.umass.cs.mallet.base.fst.Transducer

Transducer.BeamLattice, Transducer.Lattice, Transducer.ViterbiLattice, Transducer.ViterbiPath, Transducer.ViterbiPath_NBest, Transducer.ViterbiPathBeam, Transducer.ViterbiPathBeamB, Transducer.ViterbiPathBeamFB, Transducer.ViterbiPathBeamKL

Field Summary

Fields inherited from class edu.umass.cs.mallet.base.fst.Transducer

INFINITE_COST, inputPipe, outputPipe, ZERO_COST

Constructor Summary

Constructors

Constructor and Description
HMM(Alphabet inputAlphabet, Alphabet outputAlphabet)
HMM(Pipe inputPipe, Pipe outputPipe)

Method Summary

Modifier and Type	Method and Description
void	addFullyConnectedStates(String[] stateNames)
void	addFullyConnectedStatesForBiLabels()
void	addFullyConnectedStatesForLabels()
void	addFullyConnectedStatesForThreeQuarterLabels(InstanceList trainingSet)
void	addFullyConnectedStatesForTriLabels()
String	addOrderNStates(InstanceList trainingSet, int[] orders, boolean[] defaults, String start, Pattern forbidden, Pattern allowed, boolean fullyConnected) Assumes that the HMM's output alphabet contains Strings.
void	addSelfTransitioningStateForAllLabels(String name)
void	addState(String name, double initialCost, double finalCost, String[] destinationNames, String[] labelNames)
void	addState(String name, String[] destinationNames)
void	addStatesForBiLabelsConnectedAsIn(InstanceList trainingSet) Add states to create a second-order Markov model on labels, adding only those transitions the occur in the given trainingSet.
void	addStatesForHalfLabelsConnectedAsIn(InstanceList trainingSet) Add as many states as there are labels, but don't create separate weights for each source-destination pair of states.
void	addStatesForLabelsConnectedAsIn(InstanceList trainingSet) Add states to create a first-order Markov model on labels, adding only those transitions the occur in the given trainingSet.
void	addStatesForThreeQuarterLabelsConnectedAsIn(InstanceList trainingSet) Add as many states as there are labels, but don't create separate observational-test-weights for each source-destination pair of states---instead have all the incoming transitions to a state share the same observational-feature-test weights.
void	estimate()
Alphabet	getInputAlphabet()
Alphabet	getOutputAlphabet()
Transducer.State	getState(int index)
HMM.State	getState(String name)
Iterator	initialStateIterator()
boolean	isTrainable()
int	numStates()
void	print()
void	reset()
boolean	train(InstanceList ilist)
boolean	train(InstanceList ilist, InstanceList validation, InstanceList testing)
boolean	train(InstanceList ilist, InstanceList validation, InstanceList testing, TransducerEvaluator eval)
void	write(File f)

Methods inherited from class edu.umass.cs.mallet.base.fst.Transducer

averageTokenAccuracy, averageTokenAccuracy, canIterateAllTransitions, forwardBackward, forwardBackward, forwardBackward, forwardBackward, forwardBackward, forwardBackward, forwardBackward, forwardBackward, forwardBackwardBeam, forwardBackwardBeam, forwardBackwardBeam, forwardBackwardBeam, forwardBackwardBeam, forwardBackwardBeam, forwardBackwardBeam, forwardBackwardBeam, generatePath, getBeamWidth, getInputPipe, getNstatesExpl, getOutputPipe, getViterbiLattice, incIter, isGenerative, pipe, setBeamWidth, setCurIter, setKLeps, setRmin, setTrainable, setUseForwardBackwardBeam, stateIndexOfString, sumNegLogProb, transduce, viterbiPath_NBest, viterbiPath_NBest, viterbiPath, viterbiPath, viterbiPath, viterbiPathBeam, viterbiPathBeam, viterbiPathBeam, viterbiPathBeamB, viterbiPathBeamB, viterbiPathBeamB, viterbiPathBeamB, viterbiPathBeamFB, viterbiPathBeamFB, viterbiPathBeamFB, viterbiPathBeamFB, viterbiPathBeamKL, viterbiPathBeamKL, viterbiPathBeamKL

Methods inherited from class java.lang.Object

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

Constructor Detail

HMM

public HMM(Pipe inputPipe,
           Pipe outputPipe)

HMM

public HMM(Alphabet inputAlphabet,
           Alphabet outputAlphabet)

Method Detail

getInputAlphabet

public Alphabet getInputAlphabet()

getOutputAlphabet

public Alphabet getOutputAlphabet()

print

public void print()

Overrides:

print in class Transducer

addState

public void addState(String name,
                     double initialCost,
                     double finalCost,
                     String[] destinationNames,
                     String[] labelNames)

addState

public void addState(String name,
                     String[] destinationNames)

addFullyConnectedStates

public void addFullyConnectedStates(String[] stateNames)

addFullyConnectedStatesForLabels

public void addFullyConnectedStatesForLabels()

addStatesForLabelsConnectedAsIn

public void addStatesForLabelsConnectedAsIn(InstanceList trainingSet)

Add states to create a first-order Markov model on labels, adding only those transitions the occur in the given trainingSet.

addStatesForHalfLabelsConnectedAsIn

public void addStatesForHalfLabelsConnectedAsIn(InstanceList trainingSet)

Add as many states as there are labels, but don't create separate weights for each source-destination pair of states. Instead have all the incoming transitions to a state share the same weights.

addStatesForThreeQuarterLabelsConnectedAsIn

public void addStatesForThreeQuarterLabelsConnectedAsIn(InstanceList trainingSet)

Add as many states as there are labels, but don't create separate observational-test-weights for each source-destination pair of states---instead have all the incoming transitions to a state share the same observational-feature-test weights. However, do create separate default feature for each transition, (which acts as an HMM-style transition probability).

addFullyConnectedStatesForThreeQuarterLabels

public void addFullyConnectedStatesForThreeQuarterLabels(InstanceList trainingSet)

addFullyConnectedStatesForBiLabels

public void addFullyConnectedStatesForBiLabels()

addStatesForBiLabelsConnectedAsIn

public void addStatesForBiLabelsConnectedAsIn(InstanceList trainingSet)

Add states to create a second-order Markov model on labels, adding only those transitions the occur in the given trainingSet.

addFullyConnectedStatesForTriLabels

public void addFullyConnectedStatesForTriLabels()

addSelfTransitioningStateForAllLabels

public void addSelfTransitioningStateForAllLabels(String name)

addOrderNStates

public String addOrderNStates(InstanceList trainingSet,
                              int[] orders,
                              boolean[] defaults,
                              String start,
                              Pattern forbidden,
                              Pattern allowed,
                              boolean fullyConnected)

Assumes that the HMM's output alphabet contains Strings. Creates an order-n HMM with input predicates and output labels given by trainingSet and order, connectivity, and weights given by the remaining arguments.

Parameters:

trainingSet - the training instances

orders - an array of increasing non-negative numbers giving the orders of the features for this HMM. The largest number n is the Markov order of the HMM. States are n-tuples of output labels. Each of the other numbers k in orders represents a weight set shared by all destination states whose last (most recent) k labels agree. If orders is null, an order-0 HMM is built.

defaults - If non-null, it must be the same length as orders, with true positions indicating that the weight set for the corresponding order contains only the weight for a default feature; otherwise, the weight set has weights for all features built from input predicates.

start - The label that represents the context of the start of a sequence. It may be also used for sequence labels.

forbidden - If non-null, specifies what pairs of successive labels are not allowed, both for constructing norder states or for transitions. A label pair (u,v) is not allowed if u + "," + v matches forbidden.

allowed - If non-null, specifies what pairs of successive labels are allowed, both for constructing norder states or for transitions. A label pair (u,v) is allowed only if u + "," + v matches allowed.

fullyConnected - Whether to include all allowed transitions, even those not occurring in trainingSet,

getState

public HMM.State getState(String name)

numStates

public int numStates()

Specified by:

numStates in class Transducer

getState

public Transducer.State getState(int index)

Specified by:

getState in class Transducer

initialStateIterator

public Iterator initialStateIterator()

Specified by:

initialStateIterator in class Transducer

isTrainable

public boolean isTrainable()

Overrides:

isTrainable in class Transducer

reset

public void reset()

estimate

public void estimate()

train

public boolean train(InstanceList ilist)

Overrides:

train in class Transducer

train

public boolean train(InstanceList ilist,
                     InstanceList validation,
                     InstanceList testing)

train

public boolean train(InstanceList ilist,
                     InstanceList validation,
                     InstanceList testing,
                     TransducerEvaluator eval)

write

public void write(File f)