Abstract
Restaurant owners must reliably assess restaurant customers in order to function effectively and productively to enhance the restaurant's service. It is important to have a successful forecast in order to prevent losses and boost service and market optimization. There are a variety of machine learning (ML) approaches that can be used to make these predictions, but each visitor is unique and will act in a unique way. As a result, we want to estimate how many guests a restaurant may expect in the future using big data and supervised training in this study. We used three different machine learning methods in a real dataset from supervised training to predict how many visitors a restaurant dataset "Recruit restaurant visitor forecasting" will receive: Neural Network, XGBoost and Random Forest regressor. The predicted values were compared to the real data after the simulation. Basically, algorithms used had mean errors of less than 9.5278, but the Random Forest regressor exceeded, with mean errors of 9.2902.