A hybrid human dynamics model on analyzing hotspots in social networks (Q1925519)

Summary: The increasing development of social networks provides a unique source for analyzing human dynamics in the modern age. In this paper, we analyze the top-one Internet forum in China (``Tianya Club'') and identify the statistical properties of hotspots, which can promptly reflect the crowd events in people's real-life. Empirical observations indicate that the interhotspot distribution follows a power law. To further understand the mechanism of such dynamic phenomena, we propose a hybrid human dynamic model that combines ``memory'' of individual and ``interaction'' among people. To build a rich simulation and evaluate this hybrid model, we apply three different network datasets (i.e., WS network, BA network, and Karate-Club). Our simulation results are consistent with the empirical studies, which indicate that the model can provide a good understanding of the dynamic mechanism of crowd events using such social networking data. We additionally analyze the sensitivity of model parameters and find the optimal model settings.