Learning Sentimental Influence in Twitter

Recently, research about social networks has attracted tremendous interests. It can be considered that the links of online social networks describe the relationships between individuals. Analyzing online data from social networks provides opportunities for extracting attributes of sentimental influence, which also helps to get over the corner of current research on sentiment analysis. In this paper we design models to learn both sentimental influencing probabilities and influenced probabilities for users of Twitter, one of the most popular online social media. We find that there is a high correlation between Twitter users' influencing probabilities and influenced probabilities, and the majority of users keep sentimental balance on both.