ALCMean’s: Unsupervised community detection using local Laplacian, automatic detection of the number of centers

Community detection is a fundamental problem in the analysis of complex networks. It has applications across social, biological, and financial domains. Traditional algorithms, such as Louvain, LPA, and modularity optimization, often require manual parameter tuning. They also suffer from inaccurate cluster center selection. To address these challenges, we propose AlCMean’s, a novel algorithm. AlCMean’s combines Laplacian energy–based automatic center identification with DeepWalk embeddings for robust node representation. Unlike existing Laplacian-based and clustering methods, AlCMean’s eliminates the need to predefine the number of communities, enhances cluster center selection using structural importance, and leverages representation learning for more accurate and stable assignments. Experimental results on benchmark datasets demonstrate ۱۰–۲۰% higher NMI and ARI scores compared to Louvain, Newman–Girvan, LPA, Fast-Greedy, and a recent GNN-based competitor (MAGI, KDD’۲۴). Additional evaluations with modularity and F۱-scores confirm the superiority of AlCMean’s. Ablation studies highlight the critical contributions of each component. Despite its reliance on DeepWalk parameters and increased runtime relative to lightweight heuristics, AlCMean’s consistently outperforms state-of-the-art methods. This makes it a promising tool for real-world network analysis. The source code and datasets are publicly available at https://github.com/shahinmomenzadeh/ALCMeans.git

کلیدواژه ها:

Complex Networks ، Automatic Cluster Center Selection ، Social Network Analysis ، Graph Theory