com.hankcs.hanlp.dependency.perceptron.parser

Class KBeamArcEagerDependencyParser

java.lang.Object

com.hankcs.hanlp.dependency.AbstractDependencyParser

com.hankcs.hanlp.dependency.perceptron.parser.KBeamArcEagerDependencyParser

All Implemented Interfaces:

IDependencyParser

public class KBeamArcEagerDependencyParser
extends AbstractDependencyParser

基于ArcEager转移系统以平均感知机作为分类器的柱搜索依存句法分析器

Author:

hankcs

Constructor Summary

Constructors

Constructor and Description
KBeamArcEagerDependencyParser()
KBeamArcEagerDependencyParser(KBeamArcEagerParser parser)
KBeamArcEagerDependencyParser(Segment segment, KBeamArcEagerParser parser)
KBeamArcEagerDependencyParser(String modelPath)
KBeamArcEagerDependencyParser(String modelPath, String cwsModelPath, String posModelPath)

Method Summary

Modifier and Type	Method and Description
double[]	evaluate(String testCorpus) 标准化评测
CoNLLSentence	parse(List<Term> termList) 分析句子的依存句法
CoNLLSentence	parse(List<Term> termList, int beamWidth, int numOfThreads) 执行句法分析
static KBeamArcEagerDependencyParser	train(String trainCorpus, String devCorpus, String clusterPath, String modelPath) 训练依存句法分析器

Methods inherited from class com.hankcs.hanlp.dependency.AbstractDependencyParser

enableDeprelTranslator, getDeprelTranslator, getSegment, parse, setDeprelTranslater, setDeprelTranslator, setSegment

Methods inherited from class java.lang.Object

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

Constructor Detail

KBeamArcEagerDependencyParser

public KBeamArcEagerDependencyParser()
                              throws IOException,
                                     ClassNotFoundException

Throws:

IOException

ClassNotFoundException

KBeamArcEagerDependencyParser

public KBeamArcEagerDependencyParser(Segment segment,
                                     KBeamArcEagerParser parser)

KBeamArcEagerDependencyParser

public KBeamArcEagerDependencyParser(KBeamArcEagerParser parser)

KBeamArcEagerDependencyParser

public KBeamArcEagerDependencyParser(String modelPath)
                              throws IOException,
                                     ClassNotFoundException

Throws:

IOException

ClassNotFoundException

KBeamArcEagerDependencyParser

public KBeamArcEagerDependencyParser(String modelPath,
                                     String cwsModelPath,
                                     String posModelPath)
                              throws IOException,
                                     ClassNotFoundException

Throws:

IOException

ClassNotFoundException

Method Detail

train

public static KBeamArcEagerDependencyParser train(String trainCorpus,
                                                  String devCorpus,
                                                  String clusterPath,
                                                  String modelPath)
                                           throws InterruptedException,
                                                  ExecutionException,
                                                  IOException,
                                                  ClassNotFoundException

训练依存句法分析器

Parameters:

trainCorpus - 训练集

devCorpus - 开发集

clusterPath - Brown词聚类文件

modelPath - 模型储存路径

Throws:

InterruptedException

ExecutionException

IOException

ClassNotFoundException

evaluate

public double[] evaluate(String testCorpus)
                  throws IOException,
                         ExecutionException,
                         InterruptedException

标准化评测

Parameters:

testCorpus - 测试语料

Returns:

包含UF、LF的数组

Throws:

IOException

ExecutionException

InterruptedException

parse

public CoNLLSentence parse(List<Term> termList)

Description copied from interface: IDependencyParser

分析句子的依存句法

Parameters:

termList - 句子，可以是任何具有词性标注功能的分词器的分词结果

Returns:

CoNLL格式的依存句法树

parse

public CoNLLSentence parse(List<Term> termList,
                           int beamWidth,
                           int numOfThreads)

执行句法分析

Parameters:

termList - 分词结果

beamWidth - 柱搜索宽度

numOfThreads - 多线程数

Returns:

句法树