Analysis and Prediction of Short-Term Subway Passenger Flow Characteristics Based on LSTM Model

DOI

10.2991/978-94-6463-986-5_57How to use a DOI?

Keywords

Long Short-Term Memory; Passenger flow forecast; Deep learning

Abstract

With the deepening urbanization process, metro systems have become one of the important components of urban passenger transportation. However, the growing passenger flow poses severe operational challenges, giving rise to the demand for short-term passenger flow prediction (STPFP). This study employs a Long Short-Term Memory (LSTM) network model to analyze and predict short-term metro passenger flow based on current data. The dateset includes Beijing Metro ridership records from 2016, along with meteorological and air quality data. Multiple data multiprocessing steps—including outlier detection (IQR method) and data standardization (min-max scaling)—were applied to optimize the dateset. A three-layer bidirectional LSTM network was constructed, incorporating Dropout layers to prevent over-fitting. Higher weight was assigned to prediction errors during peak morning and evening hours. Subsequent optimization and validation were conducted based on evaluation metrics. By adjusting parameters such as neuron count, learning rate, and batch size, the model’s ability to capture sequential features was enhanced. The results demonstrate significant performance improvements: MSE decreased by 53.5%, RMSE by 31.5%, MAE by 31.7%, and R2 increased by 18.5% to 0.8801, with notably improved prediction accuracy during peak hours. These findings can provide decision-making support for traffic management and resource allocation.

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  - Haoyu Wu
PY  - 2026
DA  - 2026/02/18
TI  - Analysis and Prediction of Short-Term Subway Passenger Flow Characteristics Based on LSTM Model
BT  - Proceedings of the 2025 International Conference on Electronics, Electrical and Grid Technology (ICEEGT 2025)
PB  - Atlantis Press
SP  - 555
EP  - 568
SN  - 2352-5401
UR  - https://doi.org/10.2991/978-94-6463-986-5_57
DO  - 10.2991/978-94-6463-986-5_57
ID  - Wu2026
ER  -