Predicting Public Transport Passenger Using Machine Learning Algorithms

Meza A.; Acuna A.; Santisteban J. · 10th International Conference on Digital Arts, Media and Technology, DAMT 2025 and 8th ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering, NCON 2025 (2025) · DOI: 10.1109/ectidamtncon64748.2025.10962052

Resumen

This study compares five predictive models for passenger demand at the Lima Metro's Line 1 stations: Prophet, Linear Regression, Random Forest, Decision Tree, and AutoRegressive Integrated Moving Average (ARIMA). The analysis utilizes a dataset extracted from the OSITRAN website, which includes information on the number of passengers who used the metro at various stations during specific times. Data was collected from July 1 to July 7, 2024, between 5 a.m. and 11 p.m., covering all 26 metro stations. The evaluation metrics used in this study are Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and the coefficient of determination (R2). The results indicate that Prophet achieved the highest scores across all three metrics, establishing it as a viable solution compared to the other models analyzed.

Instituciones: Universidad Peruana de Ciencias Aplicadas