Adaptive Personalized Recommendation Systems: A systematic Review

Recommender systems assist users in navigating the vast selection of choices by offering personalized suggestions based on preferences. Originally used in e-commerce and streaming services, these systems are now applied in various sectors such as healthcare, education, and more, making them increasingly important. Despite their growth, recommender systems still face challenges, especially when addressing users whose preferences change over time.

This paper presents a review of recent research on recommender systems that deliver personalized and adaptive recommendations for users with evolving preferences. Analyzing 97 studies published between 2020 and 2024, the review categorizes them across multiple dimensions to address key research questions.

The findings reveal a diverse landscape of evaluation metrics, datasets, adaptation mechanisms, and application domains within adaptive personalized recommender systems (AdPRSs), with MovieLens as the most widely used dataset and the attention mechanism as the predominant adaptation approach. Furthermore, the review introduces a novel categorization of AdPRSs based on adaptation mechanism. By synthesizing current research, this review highlights key challenges faced in the field and identifies future directions for enhancing the efficiency and effectiveness of AdPRSs. These insights are of significant value to both practitioners and academic researchers, providing a foundation for advancing the development and optimization of AdPRSs.