Food and Beverage Recommendation in EatAja Application Using the Alternating Least Square Method Recommender System

EatAja is a startup in Indonesia that provides mobile application-based food and beverage ordering solution for restaurants. The application uses transaction data to recommend menus customers. Previous studies have developed recommender systems using the Apriori Collaborative Filtering methods. However, there are shortcomings recommendation system both methods, i.e., lack of personalization factors low scalability. learning method with matrix factorization can overcome problem. In this study, we improve product Alternating Least Square (ALS) on Apache Spark. We will compare results ALS method. comparison Mean Absolute Error (MAE) evaluation showed MAE value decreased by 0.07 Matrix