Feature Analysis of Rumor Refuting Message Commentators on Social Media

Abstract

In recent years, social networking platforms have rapidly become an important source of rumor propagation. The existence of internet rumors not only disturbs the normal thinking of the public, but also affects the security and stability of the society. This paper makes a sentiment analysis of comments on rumor refuting microblogs, to analyze the emotional tendency and characteristics of commentators from five main rumor categories: economics, society, military, politics and disaster. It finds out the factors that influence Weibo users’ positive and negative emotional tendencies. The paper shows there is a positive correlation between the emotional tendency of microblog contents and users’ emotional tendency to comments, except military category. Meanwhile, features such as gender, age, user labels, location and sentiment tendency also varies between commentators considering categories. Therefore, personalized recommendation services can be provided to target groups to trigger the spread of the truth, reduce the impact of rumors.