Recommender systems are software applications that provide product recommendations for users based on their purchase history or ratings of items. The product recommendations are likely to be of interest to the users and encompass items such as books, music CDs, movies, restaurants, documents (news articles, medical texts, and Wikipedia articles), and other services. In this paper, we propose a framework for collaborative filtering to enhance recommendation accuracy. The proposed approach summarized in two steps: (1) item-based collaborative filtering and (2) singular-value-decomposition-based collaborative filtering. In item-based collaborative filtering, the similarity between the target item and any other item is calculated. Then, the most similar items are recommended. The Singular Value Decomposition based approach handles the problem of scalability and sparsity posed by collaborative filtering and improves the performance of item-based collaborative filtering. We have tested the proposed approach by A Large-Scale Arabic Book Reviews (LABR) dataset. We used four different datasets to compare our approach with existing work. The proposed approach evaluated using the most common metrics found in the collaborative filtering: the mean absolute error (MAE) and the root mean squared error (RMSE). The proposed approach achieved high performance and obtained minimum errors in terms of RMSE and MAE values.

Collaborative filtering (CF) k-Nearest Neighbors (KNN) Item-based collaborative filtering Matrix Factorization (MF) Singular Value Decomposition (SVD) the mean absolute error (MAE) root mean squared error (RMSE).