网络文章中句子倾向性判别的方法研究  

Internet Sentence Tendency Discriminant Method

作　　者：程博文[1] 张亚航[1] 

机构地区：[1]北京大学软件与微电子学院信息安全系,北京100084

出　　处：《信息网络安全》2012年第3期14-16,共3页Netinfo Security

摘　　要：文章主要提出了一种汉语句子倾向性判别的方法,以情感词的静态极性(褒贬性)为基础,通过对其上下文的分析,通过情感修饰词的修饰关系,计算其动态(修饰)极性,从而提高了句子倾向性识别的准确性。实验结果表明,通过该种方法能够较好地识别句子的倾向性是合理有效的。This paper mainly mentions a method of how to determine the semantic orientation of sentiment words in Chinese sentences. Based on the words original semantic orientation, we figure out the words context semantic orientation with the relationship of decorated words around them. Thus we improve the accuracy for sentimental descriptive terms. The experiment has shown that the proposed approach is reasonable and effective.

关 键 词：互联网 概念 义原 倾向性 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]