TLFSL: link prediction in multilayer social networks using trustworthy Lévy-flight semi-local random walk.

The article focuses on the development and evaluation of the Trustworthy Lévy-flight Semi-Local (TLFSL) random walk framework for link prediction in multilayer social networks. TLFSL integrates intralayer and interlayer information by combining topological and multimodal features with trustworthy pathways to improve prediction accuracy, particularly in weighted and large-scale networks. It employs a semi-local random walk strategy that allows jumps to semantically similar but distant nodes, enhanced by a distributed local community detection method to reduce computational complexity. Experimental results on real-world multilayer datasets, such as the RATTUS network, demonstrate that TLFSL consistently outperforms state-of-the-art single-layer and multilayer link prediction methods in terms of precision and AUC metrics. The study highlights the importance of multilayer network structures and suggests future research directions including model refinement, integration of advanced machine learning techniques, and application to diverse social network analysis tasks.