Key points are not available for this paper at this time.
This systematic review of book recommender systems (BRS) examines relationships between recommendation techniques, algorithms, and features, and their effects on reading engagement dimensions. Following the PRISMA 2020 guidelines, 32 studies from 2020 to 2024 were analyzed using network analysis and centrality measures in four major databases. Six primary BRS development goals emerged, with technical improvement dominating (71.43%) over reading promotion (28.57%). Network analysis revealed a multipartite graph with 20 nodes and 105 edges, where Collaborative Filtering demonstrated highest centrality (weighted degree = 151) and the strongest connection with Machine Learning algorithms (weight = 24, 75% of studies). Three architectural clusters emerged: user-centric approaches centered on Collaborative Filtering, content-centric strategies integrating Hybrid Filtering with Neighborhood Methods, and specialized techniques with limited connectivity. Book-related Features (84.38%) and User-related Features (81.25%) showed highest implementation frequency. Reading engagement analysis demonstrated a significant imbalance: Behavioral Engagement appeared in 68.75% of studies, Affective Engagement in 56.25%, Cognitive Engagement in 43.75%, but Social Engagement in only 15.63%. Behavioral and Affective Engagement showed highest co-occurrence (37.50%), followed by Behavioral-Cognitive and Cognitive-Affective pairs (28.13% each). Challenges include limited integration of Social Interaction Features, algorithmic bias concerns, and insufficient consideration of cultural diversity. This study provides a systematic mapping of the relationships between the BRS components and the foundation for understanding how BRS supports reading engagement across contexts.
Malakul et al. (Sun,) studied this question.