Modeling and Predicting Retweeting Dynamics via a Mixture Process

Jinhua Gao, Huawei Shen, Shenghua Liu, Xueqi Cheng

Abstract: Modeling and predicting retweeting dynamics in social media has important implications to an array of applications. Existing models either fail to model the triggering effect of retweeting dynamics, e.g., the model based on reinforced Poisson process, or are hard to be trained using only the retweeting dynamics of individual tweet, e.g., the model based on self-exciting Hawkes process. In this paper, motivated by the observation that each retweeting dynamics is generally dominated by a handful of key nodes that separately trigger a high number of retweets, we propose a mixture process to model and predict retweeting dynamics, with each subprocess capturing the retweeting dynamics initiated by a key node. Experiments demonstrate that the proposed model outperforms the state-of-the-art model.