A Text Opinion Analysis Model for Blogosphere

摘要

Blogosphere is a collective community of lots of blogs and the related page links. Each blogger could publish the topic logs or comment the others. This paper proposes a text opinion analysis model for blogosphere. Oriented to the blogosphere, the model adopts two calculation methods for shorter blog comment and longer one. In the processing, the model calculates by summing the opinion weight of each character, constructs the text opinion case knowledge base for long text, and reuses the evaluation result of historical case to improve the efficiency. On the other side, the model analyzes the amount of topics, reviews, comments and the semantic opinion of each comment, and evaluates the influence of each blogger in blogosphere. In the experiment, we download over 80,000 blogs from the SINA (http://blog.sina.com) and RENREN (http://wwwrenren.com), construct an experimental corpus about "Blog Comment Opinion Analysis" (named BCOA) and test the precision and recall rate of the blog comment opinion analysis algorithm.