com.hankcs.hanlp.model.perceptron

Class PerceptronClassifier

java.lang.Object

com.hankcs.hanlp.model.perceptron.PerceptronClassifier

Direct Known Subclasses:

PerceptronNameGenderClassifier

public abstract class PerceptronClassifier
extends Object

感知机二分类器

Author:

hankcs

Constructor Summary

Constructors

Constructor and Description
PerceptronClassifier()
PerceptronClassifier(LinearModel model)
PerceptronClassifier(String modelPath)

Method Summary

Modifier and Type	Method and Description
protected static void	addFeature(String feature, FeatureMap featureMap, List<Integer> featureList) 向特征向量插入特征
com.hankcs.hanlp.model.perceptron.PerceptronClassifier.BinaryClassificationFMeasure	evaluate(com.hankcs.hanlp.model.perceptron.PerceptronClassifier.Instance[] instanceList) 评估
com.hankcs.hanlp.model.perceptron.PerceptronClassifier.BinaryClassificationFMeasure	evaluate(String corpus) 评估
protected abstract List<Integer>	extractFeature(String text, FeatureMap featureMap) 特征提取
LinearModel	getModel()
String	predict(String text) 预测
com.hankcs.hanlp.model.perceptron.PerceptronClassifier.BinaryClassificationFMeasure	train(String corpus, int maxIteration) 训练
com.hankcs.hanlp.model.perceptron.PerceptronClassifier.BinaryClassificationFMeasure	train(String corpus, int maxIteration, boolean averagePerceptron) 训练

Methods inherited from class java.lang.Object

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

Constructor Detail

PerceptronClassifier

public PerceptronClassifier()

PerceptronClassifier

public PerceptronClassifier(LinearModel model)

PerceptronClassifier

public PerceptronClassifier(String modelPath)
                     throws IOException

Throws:

IOException

Method Detail

train

public com.hankcs.hanlp.model.perceptron.PerceptronClassifier.BinaryClassificationFMeasure train(String corpus,
                                                                                                 int maxIteration)

训练

Parameters:

corpus - 语料库

maxIteration - 最大迭代次数

Returns:

模型在训练集上的准确率

train

public com.hankcs.hanlp.model.perceptron.PerceptronClassifier.BinaryClassificationFMeasure train(String corpus,
                                                                                                 int maxIteration,
                                                                                                 boolean averagePerceptron)

训练

Parameters:

corpus - 语料库

maxIteration - 最大迭代次数

averagePerceptron - 是否使用平均感知机算法

Returns:

模型在训练集上的准确率

predict

public String predict(String text)

预测

Parameters:

text -

Returns:

evaluate

public com.hankcs.hanlp.model.perceptron.PerceptronClassifier.BinaryClassificationFMeasure evaluate(String corpus)

评估

Parameters:

corpus -

Returns:

evaluate

public com.hankcs.hanlp.model.perceptron.PerceptronClassifier.BinaryClassificationFMeasure evaluate(com.hankcs.hanlp.model.perceptron.PerceptronClassifier.Instance[] instanceList)

评估

Parameters:

instanceList -

Returns:

extractFeature

protected abstract List<Integer> extractFeature(String text,
                                                FeatureMap featureMap)

特征提取

Parameters:

text - 文本

featureMap - 特征映射

Returns:

特征向量

addFeature

protected static void addFeature(String feature,
                                 FeatureMap featureMap,
                                 List<Integer> featureList)

向特征向量插入特征

Parameters:

feature - 特征

featureMap - 特征映射

featureList - 特征向量

getModel

public LinearModel getModel()