Short-term Subway Passenger Flow Prediction on Holiday Based on LightGBM and LSTM

DOI

10.2991/978-94-6239-648-7_8How to use a DOI?

Keywords

Subway; LightGBM; prediction; holiday

Abstract

Traditional predictive models struggle to accurately forecast short-term holiday passenger flow in urban rail transit systems. Intense fluctuations and nonlinear patterns often cause infrastructure strain, overcrowding, delays, and safety risks. Addressing this gap is vital for effective transit management, as precise forecasting enables optimized scheduling, resource allocation, and proactive crowd control. To address this issue, LightGBM and Long Short-Term Memory (LSTM) models, with feature engineering including one-hot encoded day-of-week features to distinguish holiday periods, utilize minute-level subway IC card data to predict the 15-minute inbound passenger flow at Beijing West Station during the 2019 Labor Day holiday. These models leverage 15-minute intervals and holiday-specific features. They effectively capture dynamic fluctuations such as extended morning peaks and midday rushes—patterns traditional commuter data-trained models miss. A comparative analysis evaluates their performance under the same conditions. This research provides a framework for accurate 15-minute forecasts. It enhances operational efficiency by reducing peak waiting times by 15–25 minutes, lowers overcrowding risks, and offers insights for other high-traffic hubs.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Yuhan Xia
PY  - 2026
DA  - 2026/04/24
TI  - Short-term Subway Passenger Flow Prediction on Holiday Based on LightGBM and LSTM
BT  - Proceedings of the International Workshop on Advances in Deep Learning for Image Analysis and Computer Vision (IWADIC 2025)
PB  - Atlantis Press
SP  - 60
EP  - 68
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6239-648-7_8
DO  - 10.2991/978-94-6239-648-7_8
ID  - Xia2026
ER  -