Abstract: This paper analyzes the objectivity of appraisal of content in Webpages and feasibility of appraisable classification, and introduces theory and implement method of Web text appraisable classification. Combined existing technique of feature selecting with the method of appraisable feature extracting, the thesis implementes appraisable classification in some celebrities’ Webpages by several classification algorithm, and achieved preferable effects.

Key words: Appraisable classification, Appraisable feature extracting, Classification

摘要： 分析了网页内容褒贬色彩的客观性和褒贬倾向性分类的可行性，介绍了Web文本褒贬倾向性分类的原理和实现方法，并将已有的特征选择方法与褒贬特征提取技术结合起来，使用几种分类算法实现了名人网页的褒贬倾向性分类，达到了较好的分类效果。

关键词: 褒贬倾向性分类, 褒贬特征提取, 分类

CLC Number: 

• TP391