Title:Predictability of User Behavior in Social Media: Bottom-Up v. Top-Down Modeling

Abstract:Recent work has attempted to capture the behavior of users on social media by modeling them as computational units processing information. We propose to extend this perspective by explicitly examining the predictive power of such a view. We consider a network of fifteen thousand users on Twitter over a seven week period. To evaluate the predictability of the users, we apply two contrasting modeling paradigms: computational mechanics and echo state networks. Computational mechanics seeks to construct the simplest model with the maximal predictive capability, while echo state networks relax from very complicated dynamics until predictive capability is reached. We demonstrate that the behavior of users on Twitter can be well-modeled as processes with self-feedback. We find that the two modeling approaches perform very similarly for most users, but that users where the two methods differ in performance highlight the challenges faced in applying predictive models to dynamic social data.