marmot.morph

Class MorphModel

java.lang.Object

marmot.core.Model

marmot.morph.MorphModel

All Implemented Interfaces:

java.io.Serializable

public class MorphModel
extends Model

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	MorphModel.MorphEntry

Field Summary

Fields inherited from class marmot.core.Model

BORDER_SYMBOL_, EMPTY_SYMBOL_

Constructor Summary

Constructors

Constructor and Description
MorphModel()

Method Summary

Modifier and Type	Method and Description
void	addIndexes(Word word, boolean insert)
SymbolTable<java.lang.Character>	getCharTable()
RankerModel	getLemmaModel()
boolean	getLemmaUseMorph()
boolean	getMarganlizeLemmas()
int	getMaxSignature()
int	getNumShapes()
int	getNumSubTags()
static java.util.List<java.util.List<java.lang.Integer>>	getOpenPosTagClassesCrossValidation(java.util.Collection<Sequence> sentences, int num_folds, java.util.List<SymbolTable<java.lang.String>> tag_tables)
SymbolTable<java.lang.String>	getShapeTable()
int[]	getTagCandidates(Sequence sequence, int index, State state)
int[][][]	getTagToSubTags()
SymbolTable<java.lang.String>	getTokenFeatureTable()
SymbolTable<java.lang.String>	getWeightedTokenFeatureTable()
SymbolTable<java.lang.String>	getWordTable()
boolean	hasBeenObserved(int form_index, int level, int tag_index)
void	init(MorphOptions options, java.util.Collection<Sequence> sentences)
boolean	isOOV(int form_index)
boolean	isRare(int word)
void	setLemmaCandidates(State state, boolean preprune)
void	setLemmaCandidates(Token token, State state, boolean preprune, boolean training)
void	setVerbose(boolean verbose)
static Tagger	train(MorphOptions options, java.util.Collection<Sequence> train_sentences, java.util.Collection<Sequence> test_sentences)
static Tagger	train(MorphOptions options, java.util.List<Sequence> train_sequences)
static Tagger	trainOptimal(MorphOptions options, java.util.Collection<Sequence> train_sentences, java.util.Collection<Sequence> test_sentences, java.util.List<java.lang.String> parameters, java.util.List<java.util.List<java.lang.String>> values_list, java.util.List<MorphModel.MorphEntry> results)
static Tagger	trainOptimal(MorphOptions options, java.util.Collection<Sequence> train_sentences, java.util.Collection<Sequence> test_sentences, java.lang.String parameter, java.util.Collection<java.lang.String> values, java.util.List<MorphModel.MorphEntry> results)
static Tagger	trainOptimal(MorphOptions options, java.util.List<Sequence> train_sentences, java.util.List<Sequence> test_sentences)

Methods inherited from class marmot.core.Model

getBoundaryIndex, getBoundaryState, getCategoryTable, getOptions, getOrder, getTagTables, init

Methods inherited from class java.lang.Object

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

Constructor Detail

MorphModel

public MorphModel()

Method Detail

init

public void init(MorphOptions options,
                 java.util.Collection<Sequence> sentences)

hasBeenObserved

public boolean hasBeenObserved(int form_index,
                               int level,
                               int tag_index)

getOpenPosTagClassesCrossValidation

public static java.util.List<java.util.List<java.lang.Integer>> getOpenPosTagClassesCrossValidation(java.util.Collection<Sequence> sentences,
                                                                                                    int num_folds,
                                                                                                    java.util.List<SymbolTable<java.lang.String>> tag_tables)

addIndexes

public void addIndexes(Word word,
                       boolean insert)

isRare

public boolean isRare(int word)

getWordTable

public SymbolTable<java.lang.String> getWordTable()

trainOptimal

public static Tagger trainOptimal(MorphOptions options,
                                  java.util.Collection<Sequence> train_sentences,
                                  java.util.Collection<Sequence> test_sentences,
                                  java.util.List<java.lang.String> parameters,
                                  java.util.List<java.util.List<java.lang.String>> values_list,
                                  java.util.List<MorphModel.MorphEntry> results)

trainOptimal

public static Tagger trainOptimal(MorphOptions options,
                                  java.util.Collection<Sequence> train_sentences,
                                  java.util.Collection<Sequence> test_sentences,
                                  java.lang.String parameter,
                                  java.util.Collection<java.lang.String> values,
                                  java.util.List<MorphModel.MorphEntry> results)

trainOptimal

public static Tagger trainOptimal(MorphOptions options,
                                  java.util.List<Sequence> train_sentences,
                                  java.util.List<Sequence> test_sentences)

train

public static Tagger train(MorphOptions options,
                           java.util.Collection<Sequence> train_sentences,
                           java.util.Collection<Sequence> test_sentences)

getCharTable

public SymbolTable<java.lang.Character> getCharTable()

getNumShapes

public int getNumShapes()

getShapeTable

public SymbolTable<java.lang.String> getShapeTable()

isOOV

public boolean isOOV(int form_index)

getNumSubTags

public int getNumSubTags()

getTokenFeatureTable

public SymbolTable<java.lang.String> getTokenFeatureTable()

getWeightedTokenFeatureTable

public SymbolTable<java.lang.String> getWeightedTokenFeatureTable()

getTagCandidates

public int[] getTagCandidates(Sequence sequence,
                              int index,
                              State state)

Specified by:

getTagCandidates in class Model

getTagToSubTags

public int[][][] getTagToSubTags()

setVerbose

public void setVerbose(boolean verbose)

getMaxSignature

public int getMaxSignature()

train

public static Tagger train(MorphOptions options,
                           java.util.List<Sequence> train_sequences)

setLemmaCandidates

public void setLemmaCandidates(Token token,
                               State state,
                               boolean preprune,
                               boolean training)

Specified by:

setLemmaCandidates in class Model

setLemmaCandidates

public void setLemmaCandidates(State state,
                               boolean preprune)

Specified by:

setLemmaCandidates in class Model

getLemmaModel

public RankerModel getLemmaModel()

getMarganlizeLemmas

public boolean getMarganlizeLemmas()

Specified by:

getMarganlizeLemmas in class Model

getLemmaUseMorph

public boolean getLemmaUseMorph()