Search

Recommender systems are information filtering systems that forecast user-item ratings, primarily drawing insights from extensive datasets to make personalized recommendations. Movie recommendation systems offer a mechanism for helping users connect with others who share similar interests. In many cases, traditional recommender systems struggle with accuracy when dealing with sparse data. This study focuses on addressing the issue of data sparsity and seeks to remedy it by implementing singular value decomposition collaborative filtering within a web-based movie recommendation system. In this paper, we introduce a movie recommendation system that primarily utilizes singular value decomposition (SVD) and cosine similarity to generate a list of recommendations. We improve the model by implementing a factorization approach, reducing complexity by minimizing parameters. Furthermore, our approach incorporates movie content information into the calculation of item similarities. Our proposed algorithm recommends a personalized list of the top N movies to users based on their reviewed preferences. Additionally, it includes a graphical representation of proportion of movies already watched by the user and those suggested to them. [ABSTRACT FROM AUTHOR]