多层次语义规则和表情特征下的微博情感研究

Research on microblogging sentiment analysis method based on multi-level semantic rules and expression feature

基于当前机器学习方法在中文微博文本情感分析处理过程中的复杂性、低准确率等问题,文中提出在统计情感词的基础上,结合情感影响因子和语义规则,并加入表情特征这一重要元素,优化文本情感加权计算方法,提高情感判断的准确率。通过对爬取到的原创微博数据集进行实验分析,验证了新情感分析算法的可行性。

In order to resolve the complexity and low accuracy of current machine learning methods in dealing with the sentiment analysis of Chinese micro-blog texts,it combines the influential factors of sentiment and multi-level semantic rule on the basis of collecting sentiment words,and adds the important element of expression features to optimize the weighting calculation method of micro-blogs texts,improving the precision of the sentiment classification result. By experimenting on original micro-blogs of collected data set,it verifies the feasibility and effectiveness of the new sentimental analytical algorithm.