Research On Dual Evolution Pattern Mining Model For Time-Varying Community Detection In Dynamic Networks

Nowadays, more and more groups are studying community detection. Community detection also has many practical applications in real life, such as: modeling the brain in the field of neuroscience, effectively inferring relationships and preferences between users in social networks, etc. At present, most teams use time snapshots to simulate the dynamic community network, which ignores the small error caused by the change of time between each time snapshot. Therefore, in order to simulate the influence of time on this clique network, we add two time evolution matrices corresponding to both the sum matrix and the system matrix, and propose our new evolution model: double evolutionary pattern mining model for time-varying community detection in dynamic networks. At the same time we find that finally, the experimental results on the actual network show that the proposed algorithm has higher prediction accuracy in dynamic networks compared with other algorithms.