Community Detection in Social Networks Using Graph Clustering and Centrality Measures

Abstract

The rapid growth of online social networks has transformed the manner in which individuals communicate, collaborate, and exchange information. Social networking platforms generate enormous volumes of interconnected data, creating complex network structures that are difficult to analyze using conventional statistical techniques. Graph theory has emerged as one of the most powerful mathematical frameworks for understanding the structure and behavior of social networks. Through graph-based representations, individuals are modeled as nodes while interactions among them are represented as edges. This representation facilitates the identification of hidden patterns, influential actors, and community structures within social systems.

Community detection is one of the most important problems in social network analysis. Communities represent groups of nodes that exhibit stronger internal connections compared to external connections. The identification of such communities enables researchers to understand information diffusion, social influence, opinion formation, marketing effectiveness, political mobilization, and collaborative behaviors. Various graph clustering techniques have been developed to identify communities, including modularity optimization, spectral clustering, hierarchical clustering, and Louvain methods.

Alongside community detection, centrality measures provide valuable insights into the relative importance of nodes within a network. Degree centrality, betweenness centrality, closeness centrality, and eigenvector centrality are widely used to identify influential individuals, opinion leaders, and information brokers. The integration of graph clustering methods with centrality measures offers a comprehensive framework for understanding both community structures and individual influence patterns within social networks.

This study investigates the application of graph clustering and centrality measures for community detection in social networks. The research evaluates various clustering techniques, examines the effectiveness of centrality metrics, and analyzes their combined contribution toward identifying meaningful social communities. Secondary data, simulation-based analyses, and published social network datasets are utilized to assess clustering quality, influence identification, network cohesion, and community stability.

The findings indicate that graph clustering techniques effectively uncover hidden community structures, while centrality measures successfully identify influential actors within those communities. The integration of both approaches significantly enhances social network analysis and provides valuable insights for decision-making in digital communication environments.