Model for Passenger Demand Prediction in a Public Transportation Company in Peru Using Machine Learning Algorithms

Public transportation in Peru faces significant challenges due to the lack of integration of advanced technologies, which impacts operational efficiency and user satisfaction. This issue results in inefficient resource planning and reduces the quality of the service provided. To address this situation, this study proposes a Machine Learning (ML) model for predicting passenger demand. The methodology was developed in four stages: dataset definition, data preprocessing, algorithm training, and result evaluation. The model considered five main variables (F1 to F5) to train and evaluate the performance of the algorithms Random Forest (RF), K-Nearest Neighbors (KNN), Linear Regression (LR), and Decision Tree (DT) using metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE). The results showed that the RF algorithm achieved the highest accuracy and lowest error, standing out for its ability to fit real-world data, while DT demonstrated the poorest performance, with higher variability in its predictions. These findings position RF as an effective tool for accurately predicting passenger demand.