Measuring emotion bifurcation points for individuals in social media

Abstract

Social media has been a new platform for emotion expression of individuals or groups in recent years. Millions of textual messages are constantly being generated. People with different emotion perceptions have different reactions to the same emotional event occurring in real life. However, it is hard to measure individual's emotion perception ability in both real world and social networks. This paper deals with online individual's emotion in view of complex system theory, and explores the emotion expression mechanism behind tweets. An concept of emotion bifurcation point is defined to denote the emotion perception ability and a methodological framework is proposed to measure it. Under the fundamental integration of the recognized Chinese emotion dictionaries (25,651 words included in total after reconciliation), new-born emotion words (458 in total) trained from a Sina-weibo corpus with 17 million tweets and commonly used emoticons (298 in total) as full-scale as possible, an emotion element ontology is constructed. Experimental evaluation on several certificated figures on Sina-weibo are implemented and the obtained results illustrate the reliability and validity of the proposed method.