Research Article
Machine Learning Approach for Labeling Undetected Planned Trips in Public Transport Operators
@INPROCEEDINGS{10.1007/978-3-031-86370-7_11, author={Mohammad Amin Zadenoori and Marco Calamai and Francesca Del Lungo and Daria Faucci and Andrea Gaffi and Lorenzo Sarti and Alessio Micheli}, title={Machine Learning Approach for Labeling Undetected Planned Trips in Public Transport Operators}, proceedings={Intelligent Transport Systems. 8th International Conference, INTSYS 2024, Pisa, Italy, December 5--6, 2024, Revised Selected Papers}, proceedings_a={INTSYS}, year={2025}, month={4}, keywords={Machine Learning Public Transport Trip Classification Random Forest Model Deployment}, doi={10.1007/978-3-031-86370-7_11} }- Mohammad Amin Zadenoori
Marco Calamai
Francesca Del Lungo
Daria Faucci
Andrea Gaffi
Lorenzo Sarti
Alessio Micheli
Year: 2025
Machine Learning Approach for Labeling Undetected Planned Trips in Public Transport Operators
INTSYS
Springer
DOI: 10.1007/978-3-031-86370-7_11
Abstract
Accurate labeling of undetected trips in public transportation is critical, as it directly affects operational efficiency, cost savings, and service quality. Undetected trips refer to scheduled trips that were either not completed or inaccurately recorded by Automatic Vehicle Location (AVL) systems. These discrepancies can disrupt resource allocation, hinder operational planning, and compromise financial accountability. If undetected trips are not properly classified, they can cause significant financial losses, misallocation of resources, and lower customer satisfaction due to unaddressed service issues.
This paper presents a machine learning approach to automate the classification of undetected trips in public transit. The model categorizes trips into three types: Operated (successfully completed trips), Lost-Deductible (missed trips within operational limits), and Lost - Non-deductible (missed trips outside operational standards and noncompensable). Automating this process enhances operational efficiency, reduces financial losses, and streamlines claim management. By replacing manual classification with AI-driven automation, transit operators can ensure faster, more accurate trip labeling, ultimately leading to optimized resource use, better decision-making, and higher service standards.
