How Much Longer?: Estimating Bus Arrival Times with Predictive Models

Abstract

Reliable bus-arrival estimates can reduce passenger uncertainty and improve travel planning in urban public-transport systems. This work studies predictive approaches for estimating arrival times from transportation data and compares methods including Linear Regression, ARIMA, Long Short-Term Memory networks, and Gated Recurrent Units. The analysis contrasts traditional statistical and machine-learning techniques with recurrent neural models designed to capture temporal dependencies. By examining their suitability for arrival-time prediction, the study contributes evidence about how predictive modeling can support more accurate real-time passenger information and improve the practical usability of public transportation services.

Publication
Proceedings of the Argentine Symposium on Artificial Intelligence and Data Science (ASAID), 54th JAIIO 2025, Buenos Aires, Argentina