Community Discovery in Social Network via Dual-Technique

Community discovery is a crucial technique for extracting knowledge and patterns from social networks. Traditional algorithms for community discovery rely on the adjacency matrix, constructed from the neighbors of each node, to partition the nodes in a graph. This approach fails to leverage the global information of the graph. In this paper, we propose a novel algorithm for community discovery in social networks that utilizes dual-technique(CD \(^2\) T). Our algorithm incorporates neighborhood rough set theory to determine the weights of neighborhood edges, and uses the shortest path to assign weights to the edges of non-adjacent nodes. By applying non-negative matrix factorization, the weight matrix is decomposed into three non-negative matrices, from which the indicator matrix is derived. We evaluate the proposed algorithm against existing methods using both real and synthetic social network datasets. The results demonstrate that our algorithm outperforms the comparison algorithms, proving its effectiveness for community discovery in social networks.