Comparative Study of Item-Based Collaborative Filtering Algorithms for Book Recommendation Systems

Shradha U. Lipane , CSMSS Polytechnic, Chhatrapati Sambhajinagar, MS, India.

Avalokiteshvara Journal of Artificial Intelligence (AJAI)
Volume 1, Issue 1, March 2025, pp. 1-9
Research Article

---

Abstract: Book recommendation systems pivotal in enhancing user experience and engagement across digital platforms, facilitating personalized content discovery in a vast sea of information and choices. This paper presents a comprehensive study on the implementation and evaluation of item-based collaborative filtering algorithms for book recommendation systems. Two prominent algorithms, k-nearest neighbor (KNN) and singular value decomposition (SVD) are used with cosine similarity as a similarity metric. This study compares these algorithms in terms of their effectiveness in recommending books that are relevant to users. The KNN algorithm identifies the nearest neighbors of a given book based on user ratings, while SVD decomposes the user-item interaction matrix to capture latent features underlying the data. Both algorithms offer unique advantages and trade-offs, which are thoroughly analyzed in this study. The evaluation metrics include precision providing insights into the accuracy and effectiveness of the recommendation models. Additionally, the computational efficiency of each algorithm is assessed to understand its scalability in real-world applications.

Keywords: Item-based collaborative filtering algorithm, k-nearest neighbor, singular value decomposition, cosine similarity.

📄 Download Full PDF