Multidimensional Insights into Recommender Systems: A Systematic Review of Evaluation Metrics and Thematic Applications

Recommender systems offer an essential tool in managing voluminous, complex data sets, providing users with personalized insights aligned to their interests. This paper delves into a comparative analysis of the three dominant architectures of recommender systems—collaborative filtering, content-based filtering, and hybrid methods—to evaluate their efficacy. This study follows a systematic Literature review approach. Our findings illustrate the unique strengths and weaknesses of each approach. Critical challenges facing traditional recommendation systems, such as data scarcity, scalability issues, and cold start problems, are also investigated. The hybrid approach was identified as the most versatile of the three architectures, combining the strengths of collaborative and content-based filtering owing to its resilience to common recommender issues; however, it comes at the cost of complexity in design, implementation, and computational resources. Alongside, the paper illuminates the progressive advancements in their applications, techniques, and utilization scope. It underscores the emerging applications of recommender systems, sparked by the opportunities inherent in applying these systems to uncharted domains. By offering a contemporary snapshot of recommender systems research, this paper anticipates future directions for this dynamic field.