Finding influential communities in networks with multiple influence types

Abstract

Recent studies on the influential community model have discovered communities that contain highly influential members. There are many types of metrics that describe the influences of objects in networks. Existing methods, however, search for influential communities based on only one influence type without comprehensively considering other influence types. In this paper, we propose an efficient influential community search method that finds the top-gamma most influential communities across multiple influence criteria. The influences are modeled as multi-dimensional vectors, where each dimension represents an influence type. To rank communities properly, we utilize the top-gamma dominating query concept for multi-dimensional point data. Extensive experiments demonstrate that the proposed method effectively finds influential communities based on multiple influence types and is orders of magnitude faster than a baseline solution.