المستخلص: |
In the domain of Network Data Processing and Data Learning, the identification of communities is crucial for comprehending the functional characteristics of networks. Overlapping community detection, which involves clusters with shared nodes, has become increasingly important in the context of real-world networks. However, there is still a requirement for additional research and the creation of innovative algorithms that consider various factors. This research proposes an updated method for perceiving the inferences within network communities. It introduces a multi-stage approach that starts by identifying seed nodes and concludes by discovering overlapping communities. The novelty lies in the use of a graph/network metric to identify significant seed nodes. The research focuses on two categories: identifying highly significant nodes based on similarity measures and recognizing cluster centers that maximize community density. The experimental outcomes validate the efficiency and scalability of the proposed methodology in identifying overlapping communities in large-scale real-world networks. Through a comparative analysis against state-of-the-art methods, the performance of the proposed approach is further confirmed. This study makes a significant contribution to the field of community detection by presenting an innovative approach that takes into account overlapping communities and integrates graph/network metrics. The results offer valuable insights into the characteristics and functional properties of networks, thereby facilitating advancements in network data processing and data learning methodologies.
|