Emotional contagion-based social sentiment mining in social networks by introducing network communities

Abstract

The rapid development of social media services has facilitated the communication of opinions through online news, blogs, microblogs, instant-messages, and so on. This article concentrates on the mining of readers' social sentiments evoked by social media materials. Existing methods are only applicable to a minority of social media like news portals with emotional voting information, while ignore the emotional contagion between writers and readers. However, incorporating such factors is challenging since the learned hidden variables would be very fuzzy (because of the short and noisy text in social networks). In this paper, we try to solve this problem by introducing a high-order network structure, i.e. communities. We first propose a new generative model called Community-Enhanced Social Sentiment Mining (CESSM), which 1) considers the emotional contagion between writers and readers to capture precise social sentiment, and 2) incorporates network communities to capture coherent topics. We then derive an inference algorithm based on Gibbs sampling. Empirical results show that, CESSM achieves significantly superior performance against the state-of-the-art techniques for text sentiment classification and interestingness in social sentiment mining.