Abstract:
In a context where complex networks play a crucial role in many fields such as friend suggestion in social networks, biology, and recommendation systems, link prediction becomes a major issue for understanding and analyzing these interconnected structures. The topic of link prediction for complex networks explores methods aimed at anticipating potential connections between entities within these networks, focusing on anticipating potential connections between nodes. The main objective of this study is to evaluate and compare the performance of widely used similarity-based link prediction methods using various datasets. A rigorous methodology, including a five-fold cross-validation process and the use of performance measures, is implemented to compare different approaches. The study's results highlight the strengths and weaknesses of different link prediction methods in complex networks, paving the way for future research. This includes the development of a new link prediction measure to improve the efficiency of algorithms in complex networks.
