A Fast Node Communication Protocol With Local Leader Selection for Community Detection in Large-Scale Complex Networks.

In the field of cybernetics, understanding the structural intricacies of complex networks is essential for advancing computational intelligence and networked systems. One of the biggest challenges in this domain is the accurate detection of community structures, which help reveal the underlying organization and functionality of these networks. However, community detection (CD) algorithms usually suffer from the accuracy-time tradeoff, where high accuracy often comes at the cost of high computational time, while faster algorithms often lose precision. Moreover, most of the current methods lack stability, ignore information flow beyond neighbors, and are limited to unweighted networks. Motivated by routing approaches, communication protocols, and principles of graph theory, this article addresses the above limitations by introducing a fast node communication protocol for CD (NCP-CD). To achieve stability, we propose a novel local leader selection (LLS) technique that identifies key nodes without requiring predefined parameters. Moreover, this article proposes a new connectivity strength (CS) metric to accurately measure the connectivity between both adjacent and nonadjacent nodes in weighted networks. Using LLS and CS, NCP-CD detects high-quality community structures by establishing a new packet exchange mechanism between nodes in a distributed architecture. This mechanism, combined with an innovative port-specific optimization technique, enables NCP-CD to achieve a fast execution speed, even in large-scale networks. Extensive experiments on both synthetic and real-world networks show that NCP-CD significantly outperforms several baseline and state-of-the-art algorithms with respect to both accuracy and computational speed.