smile.nlp.pos

Class HMMPOSTagger

java.lang.Object

smile.nlp.pos.HMMPOSTagger

All Implemented Interfaces:

java.io.Serializable, POSTagger

public class HMMPOSTagger
extends java.lang.Object
implements POSTagger, java.io.Serializable

Part-of-speech tagging with hidden Markov model.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
HMMPOSTagger() Constructor.

Method Summary

Modifier and Type	Method and Description
static HMMPOSTagger	fit(java.lang.String[][] sentences, PennTreebankPOS[][] labels) Fits an HMM POS tagger by maximum likelihood estimation.
static HMMPOSTagger	getDefault() Returns the default English POS tagger.
static void	main(java.lang.String[] argvs) Train the default model on WSJ and BROWN datasets.
static void	read(java.lang.String dir, java.util.List<java.lang.String[]> sentences, java.util.List<PennTreebankPOS[]> labels) Load training data from a corpora.
PennTreebankPOS[]	tag(java.lang.String[] sentence) Tags the sentence in the form of a sequence of words
static void	walkin(java.io.File dir, java.util.List<java.io.File> files) Recursive function to descend into the directory tree and find all the files that end with ".POS"

Methods inherited from class java.lang.Object

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

Constructor Detail

HMMPOSTagger

public HMMPOSTagger()

Constructor. Creates an empty model. For Serialization only.

Method Detail

getDefault

public static HMMPOSTagger getDefault()

Returns the default English POS tagger.

Returns:

the default English POS tagger

tag

public PennTreebankPOS[] tag(java.lang.String[] sentence)

Description copied from interface: POSTagger

Tags the sentence in the form of a sequence of words

Specified by:

tag in interface POSTagger

fit

public static HMMPOSTagger fit(java.lang.String[][] sentences,
                               PennTreebankPOS[][] labels)

Fits an HMM POS tagger by maximum likelihood estimation.

read

public static void read(java.lang.String dir,
                        java.util.List<java.lang.String[]> sentences,
                        java.util.List<PennTreebankPOS[]> labels)

Load training data from a corpora.

Parameters:

dir - a file object defining the top directory

walkin

public static void walkin(java.io.File dir,
                          java.util.List<java.io.File> files)

Recursive function to descend into the directory tree and find all the files that end with ".POS"

Parameters:

dir - a file object defining the top directory

main

public static void main(java.lang.String[] argvs)

Train the default model on WSJ and BROWN datasets.