Two-step seeding strategy in multiplex networks with inter-layer conversion cost of influence

Abstract

In this paper, we investigate the problem of influence seeding strategy in multilayer networks. In consideration of the fact that there exist inter-layer conversion costs associated with influence diffusion between layers in multiplex networks, a novel two-step seeding strategy is proposed to identify influential individuals in multiplex networks. The first step is to determine the target layer, and the second step is to identify the target seeds. Specifically, we first propose two comparable layer selection strategies, namely, multiplex betweenness centrality and multi-hop multiplex neighbors (MMNs), to determine the target layer of seeding diffusion and then construct a multiplex gravity centrality (MGC) in the manner of the gravity model to identify the influential seeds in the target layer. Subsequently, we employ a redefined independent cascade model to evaluate the effectiveness of our proposed seeding strategy by comparing it with other commonly used centrality indicators, which is validated on both synthetic and real-world network datasets. The experimental results indicate that our proposed seeding strategy can obtain greater influence coverage. In addition, parameter analysis of a neighborhood range demonstrates that MMN-based target layer selection is relatively robust, and a smaller value of a neighborhood range can enable MGC to achieve better influence performance.