Diffusion Models for Social Analysis, Influence and Learning

Abstract

Social networks are complicated by millions of users interacting and creating massive amounts of content. The problem is that any unobservable changes in network structure can result in dramatic swings in the spread of new ideas and behaviors between users. This diffusion process leads to numerous latent information that can be extracted, analyzed, and used in different applications, including market forecasting, rumor control, disease modeling, and opinion monitoring. Furthermore, mining social media big data helps to ease tracking propagated data and trends across the world. In this article, we address the study of diffusion models in social networks. We discuss three significant categories of diffusion models: contagion, social influence, and social learning models with different enhancements applied to improve performance. The aim is to study diffusion models in social networks to effectively understand how innovation and information spread over individuals and predict future trends. Also, identifying the most influential users in social networks is addressed to help spread knowledge faster and prevent harmful content like viruses or bad online behavior from spreading. Keywords—Social Network, Information Diffusion, social influence, Predictive Models, Contusion.