Online social networks have played an increasingly important role in connecting people and spreading information. It becomes urgent and necessary to understand the information diffusion process over large scale online social networks. However, due to the intricacy of human dynamics and social interactions, the vast scale of users and information, and the heterogeneity and diversity of online social networks, understanding information diffusion in social networks remains a daunting task. The PIs explore one key question in this project: how does a piece of information travel over time and space in an online social network. Similar questions have been asked and investigated for the outbreak of epidemics, but little attempt has been made in online social networks to model the diffusion of information from both temporal and spatial dimensions due to the challenges in quantitatively modeling the diffusion process in online social networks. The PIs combine the advances in data mining, graph theory, and dynamic mathematical modeling, specifically partial differential equation modeling, to characterize and predict the temporal and spatial dynamics of information diffusion.
Broader impact: This work not only builds the first extendable theoretical framework for characterizing, modeling and predicting the diffusion process of information with dynamic mathematical modeling approach, but also helps facilitate the spreading of positive information and limit the distribution of unwanted misinformation such as spam and phishing messages because it can gain an in-depth understanding of information diffusion in social networks. Moreover, this project provides an opportunity for undergraduate students to gain valuable research experiences in mathematical modeling, graph theory and large scale data collection and analysis, and to prepare them for the challenges and opportunities in the increasingly important field of online social networks.
