Research on Book Recommendation Integrating Book Category Features and User Attribute Information
Yan Chen, Eric Blancaflor, Mideth Abisado · IEEE Access · 2025
With the increasing problem of information overload, personalized book recommendation system has become the key technology to alleviate the contradiction between readers' thirst for knowledge and resources. The traditional collaborative filtering algorithm relies on sparse user-book interaction data, which has some defects such as low recommendation accuracy and obvious long tail effect. In this paper, a personalized recommendation method is proposed, which combines book category characteristics and user attribute information(BCC-UAR). By mining the deep association between user attribute information (gender, age, occupation) and book category characteristics, the embedding vector between users and books is constructed, and the recommendation results are optimized by cosine similarity calculation. The model uses embedding technology to map discrete attributes into low-dimensional continuous vectors, and generates a 200-dimensional joint representation vector of users and books through linear transformation and feature fusion. Experiments show that the model performs well in the mean square error and average absolute error, which effectively alleviates the problem of data sparsity. The research verifies the necessity of attribute information to capture users' personalized needs, and provides theoretical and practical reference for the optimization of intelligent book recommendation system.