Modeling user attitude toward controversial topics in online social media

Abstract

The increasing use of social media platforms like Twitter has attracted a large number of online users to express their attitude toward certain topics. Sentiment, opinion, and action, as three essential aspects of user attitude, have been studied separately in various existing research work. Investigating them together not only brings unique challenges but can also help better understand a user's online behavior and benefit a set of applications related to online campaign and recommender systems. In this paper, we present a computational model that estimates individual social media user's attitude toward controversial topics in terms of the three aspects and their relationships. Our model can simultaneously capture the three aspects so as to predict action and sentiment based on one's opinions. Experiments on multiple social media campaign datasets demonstrated that our attitude model can more effectively predict people's sentiment, opinion and action than approaches that treat these aspects separately.