Information Diffusion Prediction via Recurrent Cascades Convolution

Effectively predicting the size of an information cascade is critical for many applications spanning from identifying viral marketing and fake news to precise recommendation and online advertising. Traditional approaches either heavily depend on underlying diffusion models and are not optimized for popularity prediction, or use complicated hand-crafted features that cannot be easily generalized to different types of cascades. Recent generative approaches allow for understanding the spreading mechanisms, but with unsatisfactory prediction accuracy. To capture both the underlying structures governing the spread of information and inherent dependencies between re-tweeting behaviors of users, we propose a semi-supervised method, called Recurrent Cascades Convolutional Networks (CasCN), which explicitly models and predicts cascades through learning the latent representation of both structural and temporal information, without involving any other features. In contrast to the existing single, undirected and stationary Graph Convolutional Networks (GCNs), CasCN is a novel multi-directional/dynamic GCN. Our experiments conducted on real-world datasets show that CasCN significantly improves the prediction accuracy and reduces the computational cost compared to state-of-the-art approaches.

I. Introduction

Online social platforms allow their users to generate and share various contents and communicate on topics of mutual interest. Such activities facilitate fast diffusion of information and, consequently, spur the phenomenon of information cascades. The phenomenon is ubiquitous – i.e., it has been identified in various settings: paper citations [1], blogging space [2], [3], email forwarding [4], [5]; as well as in social sites (e.g., Sina Weibo [6] and Twitter [7], [8]). A body of research in various domains has focused on modeling cascades, with significant implications for a number of applications, such as marketing viral discrimination [9], influence maximization [10], [11], media advertising [12] and fake news detection [13]–[15]. Cascade prediction problem turns out to be of utmost importance since it enables controlling (or accelerating) information spreading in various scenarios.