Movie Recommendation System using ML

This document analyzes the implementation of a movie recommendation engine based on collaborative and feature- based filtering, exploring its design on machine learning principles. The system is constructed to provide adequately formulated movie recommendations based on user preferences and browsing behavior retrieved from user data. The recommendation model uses a dataset from TMDb (The Movie Database), and the system identifies similarity metrics in user interest profiles to recommend movies that users are likely to enjoy. The recommendation system was developed in Python, while Streamlit was used to design the graphical user interface, enabling users to select a movie of interest and view associated recommendations complete with their posters pulled from the TMDb API. Other important problems like the cold-start problem and system scalability are also tackled in this paper ensuring that the system functions optimally in the different scenarios users may present. The document describes the processes of data prepossessing, model training and system evaluation, showcasing that indeed user satisfaction had been enhanced. Results also suggest that the approach taken is effective in providing pertinent recommendations that are also interesting to users. This is followed by discussions of other changes that could be made regarding the project to set it towards the sustainable development goals on innovation and improvement of user engagement and experience.