مشخصات پژوهش

The detection of non-overlapping communities in complex networks is a crucial research problem because many real-world systems can be better understood by analyzing their underlying structures. Complex networks, such as social networks, biological networks, transportation systems, and information networks, often exhibit modular organization where nodes are more densely connected internally than externally. Understanding these communities is essential for both theoretical and practical purposes. For example, in social networks, identifying communities can reveal social groups, shared interests, or influential users, which is valuable for marketing, recommendation systems, and opinion analysis. In biology, detecting functional modules or protein complexes can help to understand cellular mechanisms and disease pathways, leading to better drug discovery and therapeutic interventions. In infrastructure and transportation networks, community detection can assist in identifying critical components, improving resilience, and optimizing resource allocation. Despite the importance of community detection, many existing methods suffer from limitations such as high computational complexity, sensitivity to network size, and inability to accurately detect small or densely connected local communities. Traditional global optimization methods, spectral clustering, and modularity-based algorithms often fail to scale for large networks or overlook significant local structures. Furthermore, in many applications, fast and interpretable methods are required, especially when networks are dynamically evolving or extremely large, as in social media platforms or biological interactomes. Recent studies have shown that leveraging local structures, such as dense core subgraphs, combined with node-level metrics like attraction or influence, can substantially improve both accuracy and efficiency in detecting communities. Therefore, there is a clear necessity to develop new frameworks that combine