摘要
情感分类是一项具有较大实用价值的分类技术，它可以在一定程度上解决网络评论信息杂乱的现象，方便用户准确定位所需信息。目前针对中文情感分类的研究相对较少，该文考虑将一些网络评论进行情感分类，判断一篇评论是正面还是反面。文本分类的机器学习方法较多，该文采用支持向量机的方法进行分类。该文特点在于采用具有语意倾向的词并综合其词性作为特征项．采用TF—IDF的值作为特征项权值。实验表明，用这种方法对网上的一些评论进行分类可以达到一个高的准确率。
Sentiment classification is a classification technology which has many useful applications.to a certain extend,it can solve the clutter of network reviews ,in order to facilitate the users to precisely define the necessary information. Up to now,most research of sentiment clas- sification is on English reviews,and little work has been done on Chinese reviews.In this paper, we will introduce how to apply SVM to solve sentiment classification problems. Its main target is to determine whether the reviews is positive or negative, in this paper, we will select the words which have semantic orientation as fetures, and use TF-IDF as the fetures presence vectors. This is the innovation of this article.The experimental results show that the method achieves an high accuracy rate when used to evaluate reviews from intemet.
作者
梁坤
古丽拉·阿东别克
LIANG Kun, Gulila.Altenbek （Information Science and Engineer College, XinJiang University, Ulmqi 830046, China）
出处
《电脑知识与技术》
2009年第5期3496-3498,共3页
Computer Knowledge and Technology
关键词
情感分类
语义倾向度
支持向量机
sentiment classification
semantic tendency
support vector machine
