THE DEVELOPMENT OF INTELLIGENCE BOOK RECOMMENDATION MODEL USING NEURAL COLLABORATIVE FILTERING METHOD

Abstract

The arrival of the digital era has significantly reshaped how readers discover and interact with books, diminishing the effectiveness of conventional recommendation approaches, such as bestseller rankings and expert reviews, in reflecting personalized tastes.  This study addresses the limitations of traditional methodologies, specifically the cold start and data sparsity concerns, by developing an intelligent book recommendation system that utilizes Neural Collaborative Filtering algorithms. The aim is to achieve higher recommendation accuracy by leveraging advanced techniques in user–item interaction modelling. Data is acquired from many sources, pre-processed, and evaluated using deep learning models that detect nonlinear patterns. The system's performance is evaluated using accuracy, precision, and recall scores, with a focus on mitigating cold start and data sparsity problems. The system provides reliable recommendations to existing users. Consequently, this study makes a significant contribution to the power of neural collaborative filtering in transforming customized book suggestions into the digital world.