edu.cmu.sphinx.linguist.language.ngram
Interface LanguageModel

All Superinterfaces:

Configurable

All Known Implementing Classes:

ClassBasedLanguageModel, InterpolatedLanguageModel, KeywordOptimizerModel, LargeTrigramModel, NetworkLanguageModel, SimpleNGramModel

public interface LanguageModel
extends Configurable

Represents the generic interface to an N-Gram language model.

Note that all probabilities are in LogMath log base, except as otherwise noted.

Field Summary

static java.lang.String	PROP_DICTIONARY The property specifying the dictionary to use
static java.lang.String	PROP_FORMAT The property specifying the format of the language model.
static java.lang.String	PROP_LOCATION The property specifying the location of the language model.
static java.lang.String	PROP_MAX_DEPTH The Sphinx Property specifying the maximum depth reported by the language model (from a getMaxDepth()) call.
static java.lang.String	PROP_UNIGRAM_WEIGHT The property specifying the unigram weight

Method Summary

void	allocate() Create the language model
void	deallocate() Deallocate resources allocated to this language model
int	getMaxDepth() Returns the maximum depth of the language model
float	getProbability(WordSequence wordSequence) Gets the n-gram probability of the word sequence represented by the word list
float	getSmear(WordSequence wordSequence) Gets the smear term for the given wordSequence
java.util.Set<java.lang.String>	getVocabulary() Returns the set of words in the language model.
void	start() Called before a recognition
void	stop() Called after a recognition

Methods inherited from interface edu.cmu.sphinx.util.props.Configurable

newProperties

Field Detail

PROP_FORMAT

@S4String(defaultValue="arpa")
static final java.lang.String PROP_FORMAT

The property specifying the format of the language model.

See Also:

Constant Field Values

PROP_LOCATION

@S4String(defaultValue=".")
static final java.lang.String PROP_LOCATION

The property specifying the location of the language model.

See Also:

Constant Field Values

PROP_UNIGRAM_WEIGHT

@S4Double(defaultValue=1.0)
static final java.lang.String PROP_UNIGRAM_WEIGHT

The property specifying the unigram weight

See Also:

Constant Field Values

PROP_MAX_DEPTH

@S4Integer(defaultValue=-1)
static final java.lang.String PROP_MAX_DEPTH

The Sphinx Property specifying the maximum depth reported by the language model (from a getMaxDepth()) call. If this property is set to (-1) (the default) the language model reports the implicit depth of the model. This property allows a deeper language model to be used. For instance, a trigram language model could be used as a bigram model by setting this property to 2. Note if this property is set to a value greater than the implicit depth, the implicit depth is used. Legal values for this property are 1..N and -1.

See Also:

Constant Field Values

PROP_DICTIONARY

@S4Component(type=Dictionary.class)
static final java.lang.String PROP_DICTIONARY

The property specifying the dictionary to use

See Also:

Constant Field Values

Method Detail

allocate

void allocate()
              throws java.io.IOException

Create the language model

Throws:

java.io.IOException

deallocate

void deallocate()

Deallocate resources allocated to this language model

start

void start()

Called before a recognition

stop

void stop()

Called after a recognition

getProbability

float getProbability(WordSequence wordSequence)

Gets the n-gram probability of the word sequence represented by the word list

Parameters:

wordSequence - the wordSequence

Returns:

the probability of the word sequence in LogMath log base

getSmear

float getSmear(WordSequence wordSequence)

Gets the smear term for the given wordSequence

Parameters:

wordSequence - the word sequence

Returns:

the smear term associated with this word sequence

getVocabulary

java.util.Set<java.lang.String> getVocabulary()

Returns the set of words in the language model. The set is unmodifiable.

Returns:

the unmodifiable set of words

getMaxDepth

int getMaxDepth()

Returns the maximum depth of the language model

Returns:

the maximum depth of the language model