Detecting opinion drift from Chinese web comments based on sentiment distribution computing

Abstract

Opinion drift is regarded as the change of sentiment distribution in this paper. In opinion and sentiment mining, how to detect opinion drifts and analyze their reasons, is an important problem for Web public opinion analysis. To tackle this problem, an approach of opinion drift detection for Chinese Web comment is proposed. For a comment set during a long time about a hot event, the proposed approach first determines possible drift timestamps according to the change of comment number, computes different sentiment orientations and their distributions at these timestamps, detects opinion drift according to the distribution changes, and analyzes the influences of related events occurring in the timestamps. Extensive experiments were conducted in a real comment set of Chinese forum. The results show that drift timestamps determined and opinion drifts detected correspond to the real event, so the approach proposed in this paper is feasible and effective in the application of Web public opinion analysis. © 2013 Springer-Verlag.