基于语义规则和表情加权的中文微博情感分析方法

Chinese micro-blog emotional analysis method based on semantic rules and expression weighting

针对目前中文微博情感分析方法考虑因素不全面,从而导致情感分析结果欠佳的问题,提出一种基于语义规则和表情加权的中文微博情感分析方法.该方法在使用传统情感词典分析中文微博情感倾向的基础上,在普通情感词典中融入否定词、程度副词和网络新词,根据中文微博文本独有的语言特点和句式特点,采用从词语到分句再到复句的方式对整个中文微博进行情感分析,进而使用表情加权和语义规则进行权值求和,以确定情感倾向.实验结果表明,较另外3种中文微博情感分析方法,该方法效果更显著,其平均准确率为78.4%,平均查全率为75.2%,平均F值为76.7%.

Aiming at the problem that the current Chinese micro-blog emotional analysis methods were not comprehensive,which led to poor sentiment analysis results,a Chinese micro-blog emotional analysis method based on semantic rules and expression weighting was proposed.On the basis of using traditional emotion dictionary to analyze the emotion tendency of Chinese micro-blog,negative words,degree adverbs and network neologisms were incorporated into the general emotion dictionary.According to the unique language characteristics and sentence pattern characteristics of Chinese micro-blog text,the method of emotional analysis from words to clauses and then to complex sentences was adopted to analyze the whole Chinese micro-blog.Expression weighting and semantic rules were used to perform weight summation to determine emotional tendency.The experimental results showed that compared with the other three Chinese micro-blog emotional analysis methods,the proposed method was more effective.It had an average precision rate of 78.4%,an average recall rate of 75.2%,and an average F value of 76.7%.