Passenger Flow Prediction of Shared Bicycles Based on Convolutional Long Short-Term Memory Network

View at publisher Summarise with VIRAT
Authors: Tianyi Xin
Cited by: 0
References: 10

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

In recent years, bike-sharing systems have become an integral part of urban transportation, offering residents a convenient, low-cost, and environmentally friendly travel option for short-distance trips. However, passenger flow in bike-sharing systems is highly dynamic, influenced by temporal patterns, spatial dependencies among stations, and external factors such as holidays, which brings significant challenges to accurate demand forecasting. This study proposes a shared bicycle passenger flow prediction model based on the Convolutional Long Short-Term Memory Network (k-CNN-LSTM), aiming to improve prediction accuracy by effectively integrating spatiotemporal features. Unlike traditional single time-series or spatial models, this paper innovatively combines the k-means clustering algorithm, Convolutional Neural Network (CNN), and Long Short-Term Memory (LSTM) network, and incorporates holiday information as a binary feature to capture the special flow patterns during holidays, thereby enhancing the model's adaptability to complex data. Experimental results demonstrate that k-CNN-LSTM performs excellently in predicting multiple typical sites, successfully capturing spatiotemporal correlations, and effectively addressing the flow fluctuations caused by holidays. This approach provides new methodologies and theoretical support for the intelligent scheduling of shared bicycle systems and urban traffic management.

Fields of Study & Subjects

Traffic Prediction and Management Techniques Urban Transport and Accessibility Hydrological Forecasting Using AI

Publication details

Journal / source
IEEE International Conference on Power, Electronics and Green Energy (ICPEGE)
Pages
190–194
Publication date
28 January 2026
Language
EN