com.hankcs.hanlp.classification.classifiers

Class AbstractClassifier

java.lang.Object

com.hankcs.hanlp.classification.classifiers.AbstractClassifier

All Implemented Interfaces:

IClassifier

Direct Known Subclasses:

NaiveBayesClassifier

public abstract class AbstractClassifier
extends Object
implements IClassifier

Author:

hankcs

Constructor Summary

Constructors

Constructor and Description
AbstractClassifier()

Method Summary

Modifier and Type	Method and Description
String	classify(Document document) 预测最可能的分类
String	classify(String text) 使用一个训练出来的分类器来预测分类
IClassifier	enableProbability(boolean enable) 是否归一化分值为概率
int	label(Document document) 预测最可能的分类
Map<String,Double>	predict(Document document) 预测分类
void	train(Map<String,String[]> trainingDataSet) 训练模型
void	train(String folderPath) 用UTF-8编码的语料训练模型
void	train(String folderPath, String charsetName) 训练模型

Methods inherited from class java.lang.Object

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

Methods inherited from interface com.hankcs.hanlp.classification.classifiers.IClassifier

categorize, getModel, predict, train

Constructor Detail

AbstractClassifier

public AbstractClassifier()

Method Detail

enableProbability

public IClassifier enableProbability(boolean enable)

Description copied from interface: IClassifier

是否归一化分值为概率

Specified by:

enableProbability in interface IClassifier

Returns:

classify

public String classify(String text)
                throws IllegalArgumentException,
                       IllegalStateException

使用一个训练出来的分类器来预测分类

Specified by:

classify in interface IClassifier

Parameters:

text -

Returns:

Throws:

IllegalArgumentException

IllegalStateException

classify

public String classify(Document document)
                throws IllegalArgumentException,
                       IllegalStateException

Description copied from interface: IClassifier

预测最可能的分类

Specified by:

classify in interface IClassifier

Parameters:

document - 一个结构化的文档(注意!这是一个底层数据结构,请谨慎操作)

Returns:

最可能的分类

Throws:

IllegalArgumentException

IllegalStateException

train

public void train(String folderPath,
                  String charsetName)
           throws IOException

Description copied from interface: IClassifier

训练模型

Specified by:

train in interface IClassifier

Parameters:

folderPath - 分类语料的根目录.目录必须满足如下结构:
根目录
├── 分类A
│ └── 1.txt
│ └── 2.txt
│ └── 3.txt
├── 分类B
│ └── 1.txt
│ └── ...
└── ...
文件不一定需要用数字命名,也不需要以txt作为后缀名,但一定需要是文本文件.

charsetName - 文件编码

Throws:

IOException - 任何可能的IO异常

train

public void train(Map<String,String[]> trainingDataSet)
           throws IllegalArgumentException

Description copied from interface: IClassifier

训练模型

Specified by:

train in interface IClassifier

Parameters:

trainingDataSet - 训练数据集,用Map储存.键是分类名,值是一个数组,数组中每个元素都是一篇文档的内容.

Throws:

IllegalArgumentException

train

public void train(String folderPath)
           throws IOException

Description copied from interface: IClassifier

用UTF-8编码的语料训练模型

Specified by:

train in interface IClassifier

Parameters:

folderPath - 用UTF-8编码的分类语料的根目录.目录必须满足如下结构:
根目录
├── 分类A
│ └── 1.txt
│ └── 2.txt
│ └── 3.txt
├── 分类B
│ └── 1.txt
│ └── ...
└── ...
文件不一定需要用数字命名,也不需要以txt作为后缀名,但一定需要是文本文件.

Throws:

IOException - 任何可能的IO异常

predict

public Map<String,Double> predict(Document document)

Description copied from interface: IClassifier

预测分类

Specified by:

predict in interface IClassifier

Returns:

label

public int label(Document document)
          throws IllegalArgumentException,
                 IllegalStateException

Description copied from interface: IClassifier

预测最可能的分类

Specified by:

label in interface IClassifier

Returns:

Throws:

IllegalArgumentException

IllegalStateException