Blog Sentiment Orientation Analysis Based on Dependency Parsing

Nowadays, blogs, which contain rich opinion and emotion information, have become an important platform for exchanging sentiments among netizens on the Web. Blog search provides a convenient way for these information exchanges. In many cases, users pay more attention to blogger’s opinions and sentiments about an event when searching the blogosphere, but most existing blog search engines return results based only on topics without considering sentiment orientations. In this paper, an algorithm called SOAD (sentiment orientation analysis based on syntactic dependency) is proposed for analyzing blogs’ sentiment orientation based on dependency parsing. A Chinese blog search engine prototype system is built based on the proposed algorithm, which reprocesses blog search results for sentiment analysis. Experiments show that, SOAD algorithm has more advantages, and the prototype system implements the purpose of this paper: the blog search results are returned according to blogger’s sentiment orientation.