Effective and Efficient Dense Subgraph Query in Large-Scale Social Internet of Things

Abstract

Social Internet of Things (SIoT) is the integration of social network (SN) and the Internet of Things (IOT). Community search in SIOT is an important problem beneficial to the resource/service discovery. In this article, we address the problem from the perspective of a dense subgraph query. Specifically, we propose a core-based static dense subgraph query and a graph kernel based dynamic dense subgraph query. The two algorithms consider the large scale and the time-varying nature of the SIoT, respectively. Unlike the existing works, the static method is inspired by the first-connection-last-expansion idea. Top- k neighbors of each query node are first found by a random walk. Then, all the query nodes and their top- k neighbors are connected as the core using Steiner tree expansion. The rank constraint random sampling is utilized to extend the core to a dense subgraph. Further, by leveraging a graph kernel index and identifying the updates that may affect the results, we conduct the dynamic query in an incremental update way instead of executing it from scratch. The experiments on synthetic and real datasets show that the proposed algorithms are both effective and efficient.