ESIS: Emotion-based spreader–ignorant–stifler model for information diffusion

With the proliferation of online social networks, users are willing to post messages sharing their statuses. A piece of information can include not only substantial news but also emotional expression. As messages are re-posted among users, large cascades are created and information is spread with such emotional expression. We propose an emotion-based spreader–ignorant–stifler (ESIS) model to simulate the process of information diffusion. The proposed model categorizes information cascades into fine-grained classes, and the proportion of retweets among users for one emotion as weights on edges is introduced. We conduct experiments with artificial and real social networks. The experimental results indicate that the probability of information adoption is based both on the spreading probability and retweeting strength among users. We verify the proposed model and predict the cascade size with a real-world dataset. Compared to the latest related models, i.e., the standard SIS model and the information cascade models, the proposed ESIS model demonstrates 11.8% and 16.5% performance improvements, respectively.