Assistant Professor Dr. Fatima Amer jid Almahri
Department of Information Technology, College of Computing and Information Sciences
University of Technology and Applied Sciences, Salalah, Oman
Email: [email protected]
Development of Content-Based Filtering Model for Recommendation System Using Multiple Factors related to object Preference
Abstract — As digital technology grows exponentially, the size and complexity of many websites increase accordingly. This growth results in an overload of information for users, making it difficult to find relevant content. Recommender systems can alleviate this problem by helping users navigate information overload and find items of interest. One of the most widespread recommendation methods, content-based filtering (CBF), relies on the content’s similarity and is particularly effective for handling the cold start problem. To calculate the similarities between items, CBF uses item information, represented as attributes. In this article, our main contribution is creating a recommendation system based on Content-Based Filtering (CBF) that integrates multiple criteria. We considered various user preferences, including the genres they prefer to watch, the production date of movies, and movie components. The final model combines all these criteria—genres, production date, and movie components—into a single model, a comprehensive recommendation system. In addition, we compared each model’s performance and accuracy to determine the best model. The MovieLens dataset was used for this study. The results show that the integrated model produces the highest efficiency and accuracy in content-based filtering.