一种基于混合概率模型的文本分类方法  

A Text Classification Method Based on Hybrid Probability Model

作　　者：吴新玲[1] 

机构地区：[1]广东技术师范学院计算机科学学院,广东广州510262

出　　处：《微电子学与计算机》2011年第11期133-136,共4页Microelectronics & Computer

摘　　要：提出了一种基于类的混合概率分类方法.该方法为每一类文本独立选取能代表其本质特性的主要特征,即不同类型的文本由不同的主要特征表示,并基于各类的主要特征分别为每类文本建立相应的概率分布模型,然后再根据朴素贝叶斯方法对未知类型的文本进行分类.实验结果表明:该方法简单有效且易于实现.A hybrid probabilistic classification method was proposed based on the class. The method selects the main features that can represent the essential characteristics of a class text independently for every class text. That is different class text will be expressed by different main features. Based on the main features of each class text, the probability distribution model will be established respectively for every class text. Then the unknown class label text can be classified by naive Bayesian method. The experiment results show that the proposed method is simple, effective and easy to implemen.

关 键 词：文本数据挖掘 文本分类 属性选择 概率模型 多项分布 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]