Activities information diffusion in Chinese largest recommendation social network: Patterns and generative model

摘要: Nowadays, networks play an indispensable role in social life, and have become a new advertising medium for offline activities. Previous studies of information diffusion or behavior spread over mostly focused on models analysis virtual interaction between online users, very few them focus the propagation real world activities these networks. To address this problem, we use data obtained from Chinese largest recommendation network - Douban, study how one user to another through Douban. By using cascading subgraphs trees, break whole cascade into local subgraphs. After analyzing about 1.47 million observe statistical topological characteristics Next, find size degree distributions several common patterns topology cascades. Moreover, also some other interesting discoveries, like relation number initial adopters final size, underlying influences driving behaviors. Finally, propose model that can generate cascades follow observed, validate it by empirical analysis.