Abstract

Movie recommendation systems are crucial for enriching user interactions on movie viewing platforms. This study presents an advanced movie recommendation system developed using the R programming language and the Movie Lens dataset, which contains user-provided movie ratings from 1 to 5. The system uses collaborative filtering techniques, especially user and object-based methods, to curate. personalized movie recommendations. User-based collaborative filtering detects movie preferences by evaluating user similarities through cosine similarity, while object-based collaborative filtering evaluates movie similarities. Predicted ratings for undiscovered movies are derived from analogy users or movie reviews, which form the recommendation results. Root Mean Squared Error (RMSE) estimation on different training and testing datasets confirms the commendable prediction performance, with RMSE values of 0.93 and 0.94 for user-based strategies and 0.94 for object-based strategies. This study highlights the effectiveness of the proposed recommendation system in providing personalized movie recommendations, which promises user engagement on movie streaming platforms.