Research on building chinese micro-blog semantic lexicon

Micro-blog is one of the most popular network platform for expressing opinions and suggestions by users. Therefore, micro-blog sentimental orientation is particularly important for information retrieval, information security and automatic abstracting. In this paper, we introduce method for building micro-blog semantic lexicon. Massive data of micro-blog have been downloaded and words were extracted by frequency of use, then the improved words, sentimental algorithm has been used to calculate semantic value, in addition, emoticons and network new words are essential for semantic lexicon, we selected these element and classified into positive and negative manually. Finally, we use this lexicon to classify micro-blog, and the experimental results show that we can achieve a high improvement by this semantic lexicon.