Microblogging Sentiment Analysis Using Emotional Vector

This paper proposes a new approach to analyze public moods on certain events in micro-blogging. Combining with clinical psychology, we use emotional vector rather than traditional orientation to perform sentiment analysis. The emotional vector could constantly absorb new Internet emotional words with our algorithm and has hierarchical structure so as to do multi-level analysis. Experimental evaluations show that there is a strong correlation between bursty events and public moods, and sentiment analysis could implement effectively using the emotional vector proposed in this article.