Reliable Community Search in Dynamic Networks

Abstract

Local community search is an important research topic to support complex network data analysis in various scenarios like social networks, collaboration networks, and cellular networks. The evolution of networks over time has motivated several recent studies to identify local communities from dynamic networks. However, they only utilized the aggregation of disjoint structural information to measure the quality of communities, which ignores the reliability of communities in a continuous time interval. To fill this research gap, we propose a novel (θ, k)core reliable community (CRC) model in the weighted dynamic networks, and define the problem of the most reliable community search that couples the desirable properties of connection strength, cohesive structure continuity, and the maximal member engagement. To solve this problem, we first develop an online CRC search algorithm by proposing a definition of eligible edge set and deriving the eligible edge set based pruning rules. After that, we devise a Weighted Core Forest-Index and index-based dynamic programming CRC search algorithm, which can prune a large number of insignificant intermediate results according to the maintained weight and structure information in the index, as well as the proposed upper bound properties. Finally, we conduct extensive experiments to verify the efficiency of our proposed algorithms and the effectiveness of our proposed community model on eight real datasets under different parameter settings.