Ontology-Based Recommender Systems for E-Learning and Multimedia: A Systematic Literature Review Across Domains

Abstract

The rapid expansion of digital content across various sectors has led to an overwhelming influx of data, highlighting the need for advanced recommendation systems. Traditional methods such as Collaborative Filtering (CF) and Content-Based Filtering (CBF) face limitations like data sparsity and the Cold Start problem, which affect the accuracy of recommendations. This study explores the use of ontologies in enhancing recommendation systems, aiming to overcome these challenges by providing a semantic framework for better item and user representation. A Systematic Literature Review (SLR) methodology was employed to analyze research from 2021 to 2025, focusing on the application of ontologies in e-commerce, healthcare, education, and employment. The findings demonstrate that ontologies improve recommendation relevance, diversity, and explainability, especially in addressing the Cold Start problem. However, challenges in implementation and interpretation remain. This research contributes to the field by emphasizing the potential of integrating ontologies with Knowledge Graphs (KG) and Graph Neural Networks (GNN) to create hybrid models that enhance the accuracy and transparency of recommendations, guiding future advancements in recommendation systems.