Book Recommendation System using Matrix Factorization with SVD

Abstract

The paper introduces a Book Recommendation System that leverages Matrix Factorization with Singular Value Decomposition (SVD) to address the challenge of an overwhelming number of books in online platforms. By analyzing user-item interactions within book rating datasets, the system identifies latent components representing reader preferences and book attributes. Through SVD, the approach enhances this decomposition to capture crucial latent factors. Personalized book recommendations are then generated based on a user’s preferences and estimated ratings from other users. Evaluation results illustrate the system’s effectiveness in capturing user preferences and providing more accurate and diverse recommendations.