In recent times, most of the real-world systems are heterogeneous, containing multiple kinds of nodes and edges, which can be naturally conceptualized as multilayer networks. In this paper, we develop a methodology for detecting communities in such multilayer networks. In order to do so, first, we propose a multilayer modularity index \(Q_M\) and then develop the multilayer Louvain (ML) algorithm leveraging it ( \(Q_M\) ). The proposed algorithm can simultaneously detect communities consisting of only single type, as well as multiple types of nodes (and edges). Furthermore, it is scalable and easily adaptable to complex network structures. For evaluating the performance of the proposed multilayer Louvain (ML) algorithm, we leverage synthetic networks with preplanted multilayer community structures as well as three real-world multilayer networks (Yelp, Meetup and Digg). Results show the significance of our proposed methods in discovering homogeneous as well as heterogeneous communities over multiple layers, and also highlight its ability in producing better community structures compared to competing state-of-the-art approaches.
