Web-Based Umrah Departure Quota Prediction Application UsingaMachineLearning Approach (Case Study: PT. Rahmatan BerkahWisata)

Abstract

PT. Rahmatan Berkah Wisata faces uncertainties regarding Umrah departure quotas due to capacity determination that is still conducted manually. This study aims to design a web-based quota prediction applicationutilizingMachine Learning with the Random Forest Regression algorithm, as well as to evaluate the accuracy and functionality of the system. System development employs the Rapid Application Development (RAD) method, encompassing the stages of requirement planning, user design, construction, and cutover. The research data consistsof historical pilgrim data from the 2022–2026 period, which includes the year, package name, package type, tripduration, number of pilgrims, and departure season.The results show that the Random Forest Regression model achieved a Mean Absolute Error (MAE) of 3.265, a Mean Squared Error (MSE) of 15.186, and an R-squared (R2) value of 0.707 (70.7%), demonstrating its capability to provide sufficiently accurate predictions. The model wassuccessfully implemented into a web-based application featuring Umrah data management, prediction and retraining scheduling, departure prediction, prediction reporting, and model retraining. Furthermore, functional testingusing Black Box Testing across three user roles (President Director, Operational Manager, and Pilgrim Staff) was successfully executed across all test scenarios. Based on the findings, this application can effectively serveasadecision support tool for determining Umrah departure quotas.