Effective and Efficient Steiner Maximum Path-Connected Subgraph Search in Large Social Internet of Things

Abstract

Social Internet of Things (SIoT) is extended to integrate social networks in the Internet of Things (IoT). SIoT enriches IoT, and thus resource (or service) discovery and consolidation in SIoT becomes an important and challenging problem. In this paper, we represent the SIoT as the heterogeneous information networks (HINs) with multi-typed entities and relations, and resolve this problem from the perspective of cohesive subgraph search. Specifically, given a query node $q$ , we discover a cohesive subgraph containing $q$ from a HIN, where all nodes are of the same type as $q$ and have dense relationships. Yet existing solutions cannot be applied to various HINs, and the result subgraph is not accurate and close enough due to ignoring the connectivity among nodes within the subgraph. To this end, 1) we extend the connectivity with novel meta-path-based edge-disjoint paths to HINs, and propose the $k$ -path connected component ( $k$ -PCC) to measure the cohesiveness of subgraph in HINs; 2) we model the densest connected subgraph containing $q$ as the $k$ -PCC with the maximum connectivity including $q$ , called the Steiner Maximum Path-Connected Subgraph (SMPCS); 3) we develop efficient algorithms based on an index tree for searching the SMPCS. Extensive experiments on four real HINs are conducted to demonstrate the effectiveness and efficiency of our proposed approaches.