Causal Inference in Social Media Using Convergent Cross Mapping

Abstract

Revealing underlying causal structure in social media is critical to understanding how users interact, on which a lot of security intelligence applications can be built. Existing causal inference methods for social media usually rely on limited explicit causal context, pre-assume certain user interaction model, or neglect the nonlinear nature of social interaction, which could lead to bias estimations of causality. Inspired from recent advance in causality detection in complex ecosystems, we propose to take advantage of a novel nonlinear state space reconstruction based approach, namely Convergent Cross Mapping, to perform causal inference in social media. Experimental results on real world social media datasets show the effectiveness of the proposed method in causal inference and user behavior prediction in social media.