Clustering the Digital Footprint: Revealing User Interaction Communities

Abstract

The proliferation of social networking sites (SNS) and the expansion of the internet have enabled seamless communication among individuals on a unified platform. A graph with nodes and edges connecting the nodes might represent a social network. The nodes symbolize individ uals or entities, while the edges illustrate their interactions. Individu als who associate inside social networks and share analogous decisions, tastes, and preferences create virtual clusters or communities. Identify ing these communities can be advantageous for several objectives, such as discovering a common study domain in collaborative networks, iden tifying a target audience for marketing and recommendations, and map ping protein interaction networks inside biological systems. Community identification serves as an effective instrument for discerning intricate networks. Diverse methodologies have been suggested for community detection, each addressing the issue from a distinct viewpoint. Conse quently, extensive graph-processing community recognition techniques have become essential due to the emergence of vast and intricate net works across several areas. This research presents an innovative method for community detection that integrates node space similarity and uti lizes local knowledge. We utilize eigenvector centrality and proximity metrics to improve community detection in social networks. Compre hensive studies on both synthetic and real-world networks demonstrate the effectiveness of the suggested hybrid paradigm. The results demon strate that the hybrid technique is more effective in large-scale graphs compared to other established algorithms, exhibiting significant robust ness and efficiency.