Research On Sentiment Analysis Of Microblog

People can directly and quickly express their attitude to emergencies, public figures and hot products through microblog. The results of sentiment analysis can be used for public opinion monitoring, microblog marketing, word of mouth investigation etc, at the same time provide decision-making basis for government departments or enterprises.In this paper, we study the method of microblog sentiment analysis, and establish a system of microblog analysis. First of all, this paper reviews the current research of microblog analysis field, and introduces the main methods, including the use of the API provided by weibo and web crawler technology to get microblog data and use of natural language processing technology to preprocess the data etc, then screens and sorts out the current emotional vocabulary resources to build the emotional dictionary for Chinese microblog. Then, according to the pluralistic characteristics of microblog expression, and considering the influence of modified words and conjunction on the microblog text, therefore this paper puts forward different methods of calculating the emotion tendency for microblog emotional tendency in the three situations. The situations are only the emotional words and emoticons, adverbs of degree or negative, and conjunction. At last, the paper develops a set of system of microblog sentiment analysis, which is intuitive, interactive, dynamic and visual transmission. And we summarize and prospect the work of this paper.