MOVIE RECOMMENDATION USINGDEEPLEARNING (DL)

Abstract

With the rapid expansion of streaming platforms, users are exposed to an overwhelming number of movies, making personalized content recommendation essential. Traditional recommendation systems rely heavily on collaborative filtering or content-based filtering, but these approaches often struggle with issues such as data sparsity, limited feature representation, and lack of contextual understanding. This paper proposes a Deep Learning–based Movie Recommendation System that leverages neural networks to capture complex user–item relationships and generate accurate predictions. The model integrates embedding layers, deep neural networks, and hybrid learning techniques to analyze user history, movie metadata, and contextual preferences. A combination of user embeddings, movie embeddings, and nonlinear interaction layers enhances recommendation quality. The system was evaluated using datasets containing user ratings, genres, keywords, and cast details. Experimental results show that the deep learning model significantly outperformed classical methods, achieving lower RMSE and higher precision scores. The framework demonstrates strong generalization, scalability, and adaptability to real-world streaming environments. This work highlights the importance of deep learning in modern recommendation systems and shows how advanced feature representation can deliver more personalized, relevant movie suggestions.