Movie Recommendation System for Cold-Start Problem Using User's Demographic Data

with the exponential growth of online movies, effective recommender systems have become crucial for aiding users in discovering relevant content. However, the surge in new users and movies has led to the cold-start problem, where accurate recommendations are hindered by insufficient data. In this study, we propose the Enhanced scaled-CER model, which integrates user metadata with the MovieLens-10M dataset to address this issue. Bayesian optimization techniques are employed to optimize the model's parameters, resulting in enhanced performance. Experiments conducted on the MovieLens-10M dataset demonstrate impressive achievements, including an item coverage rate of over 95% and increased Shannon entropy, indicating higher diversity and relevance in recommended movie selections. Our research offers a comprehensive and diverse set of movie recommendations to users, even in the face of limited data availability, presenting a robust solution for the cold-start problem in movie recommendation systems.