Classification of Micro-blog Sentiment Based on Naive Bayesian Classifier

This paper is to conduct popular micro-blog for sentiment classification. The Naive Bayesian Classifier is the key in this paper, and study on pretreatment of the text of micro-blog, constructing sentiment dictionary, feature selection, feature weights and expression vector, comes up with some points and conducts the experiment. And the performance of “emoticons + twice sentiment feature extraction + BOOL” is the best pretreatment method. And this experiment gains a relatively satisfactory result.