Recommender systems have become a fundamental technology for addressing information overload in large-scale digital environments. This seminar provides a structured study of the current state of recommender systems, covering foundational concepts, major methodological approaches, an‎d influential research developments. The report examines classical techniques such as collaborative filtering an‎d matrix factorization, as well as advanced approaches including graph neural network-based models, explainable recommender systems, an‎d federated learning frameworks. A comparative analysis is conducted to eva‎luate these architectures in terms of scalability, sparsity han‎dling, explainability, an‎d privacy preservation. The study highlights the evolution of recommender systems from accuracy-oriented models toward transparent, fair, an‎d privacy-aware intelligent systems

2/18/2026 12:00:00 AM

89746

1404/12/02

The Study of the Current State of Recommender Systems

Recommender Systems Evolution , Collaborative Filtering Architectures , Matrix Factorization Models , Graph Neural Networks (GNNs) for Recommendation , Explainable Recommender Systems (XRS)