Design and Implementation of a Movie Personalized Recommendation System based on ALS Algorithm
Abstract
As big data continues to expand, personalized recommendation systems help users filter massive volumes of movie‑related data and find films matching personal preferences. In this project, we build a reproducible movie recommendation system. Experiments are carried out on the MovieLens dataset. We train an ALS collaborative‑filtering model with PySpark and tune hyperparameters through grid search. RMSE serves as the evaluation metric for prediction quality. Our experiment obtains an RMSE of 0.8667 under this setting. This outcome indicates that the model can generate reasonably reliable personalized recommendations for the given dataset.