Link Prediction in Social Networks: A Review

Link prediction in social networks has garnered significant attention due to its potential applications in various domains. This review paper synthesizes and evaluates recent advancements in link prediction methodologies across diverse research studies. It provides a comprehensive overview of the strategies proposed by different authors, emphasizing key methodologies, metrics, and their advantages. The review encompasses proximity-based algorithms, dynamic user relationship analysis, node similarity-based approaches, multiplex network methods, temporal node similarity weighting, centrality measures, negative link prediction techniques, machine learning-based comparisons, and temporal preferential attachment models. Each methodology is dissected to elucidate its core principles and effectiveness in predicting links within social networks. Furthermore, the paper discusses the experimental setups, evaluation metrics, and comparative analyses employed by researchers to validate their proposed methodologies. Through this comprehensive examination, the review paper aims to offer insights into the evolving landscape of link prediction techniques and guide future research directions in social network analysis.