Common-Truss-Based Community Search on Multilayer Graphs

Xudong Liu, Zhaonian Zou

Abstract: Multilayer graphs have a multitude of applications in diverse fields, ranging from biology to social networks. In the context of the community search in multilayer graphs, the goal is to identify a target multilayer graph that contains query vertices. However, existing algorithms do not succeed in finding a common graph in subgraphs, while also ensuring that the multilayer graph is sufficiently dense. To address this limitation, we propose the CTruss model, a novel structure for multilayer graphs. This structure aims to find a common graph that exists on each layer in the subgraph. Next, we present a Baseline algorithm to find a multilayer graph containing the query vertices. We also propose Top-Down and Bottom-Up algorithms that can compute the CTruss quickly under different conditions. Finally, we conduct experiments to demonstrate that our proposed model can achieve a higher density in some cases compared to the state of art models such as FirmTruss and K-core.