A Bibliometric Map of the Hybrid Recommender Systems Field (2003-2024): Trends, Network Structures and Future Directions

Özet

This study comprehensively examines the approximately twenty-year development of the hybrid recommender
systems field using bibliometric analysis methods. Academic publications produced between 2003 and 2024 were
compiled from various databases, and the structural dynamics of the field were revealed through citation analysis,
co-authorship networks, keyword co-occurrence, and thematic clustering techniques. Overall, the study
systematically maps the evolution of the hybrid recommender systems literature and provides a strategic roadmap
for researchers and practitioners. The findings indicate that hybrid recommender systems have gained significant
momentum, particularly over the past decade. Initially emerging from the integration of content-based and
collaborative filtering approaches, the field has evolved into more complex and powerful models through the
incorporation of deep learning, big data analytics, and artificial intelligence techniques. Moreover, these systems
are widely applied across various domains, including e-commerce, media platforms, and educational technologies.
Network analyses reveal that the field is concentrated around specific countries and leading research institutions,
while also highlighting the increasing trend of interdisciplinary collaboration. Keyword analysis further shows that
themes such as “deep learning,” “context-aware systems,” “cold-start problem,” and “explainable AI” have
become prominent in recent years. The article emphasizes that future research in hybrid recommender systems
should focus more on explainability, ethical considerations, data privacy, and user-centered design. Additionally,
the integration of heterogeneous data sources and the development of real-time recommendation mechanisms are
identified as key directions for future research.