Sentiment Analysis in Chinese BBS

摘要

With the rapid growth of the web text data, BBS (Bulletin Broad System) has already been the popular discussion forum for people exchanging their thinking and mind. If we can mine and analyze these online review data posted by the users, these data can help better understanding of public opinions greatly and plays an important role in government or enterprise's construction, especially in supporting for proper decision-making. In our study, we focus on building attitudinal words weight dictionary which is composed of 1342 words, and introduce the negative words dictionary, degree words dictionary and interjection words dictionary defined by ourselves. Then we create emotion weight calculator to calculate the emotional index of each post and classify the sentence as one of three opinion groups, positive, negative and neutral. The subject is the post we crawled from the famous Chinese BBS social network at random. The result shows that this sentiment analysis method is very effective and inspirational.