Title: A fusion model of gated recurrent unit and convolutional neural network for online ride-hailing demand forecasting

Authors: Xijin Cui; Mingxia Huang; Lei Shi

Addresses: School of Transportation and Geomatics Engineering, Shenyang Jianzhu University, 25 Hunnan Middle Road, Hunnan District, Shenyang 110168, Liaoning, China ' School of Transportation and Geomatics Engineering, Shenyang Jianzhu University, 25 Hunnan Middle Road, Hunnan District, Shenyang 110168, Liaoning, China ' China Railway Jinan Group Co., Ltd., China State Railway Group Co., Ltd., 6 Qilu Road, Huaiyin District, Jinan 250001, Shandong, China

Abstract: This paper collects and analyses the impact of weather, air quality and point of interest data on residents' daily travel, establishes a fusion model combined the convolutional neural network based on point of interest data and gated recurrent neural network prediction model to investigate the influence of weather and air quality on the demand for online ride-hailing, uses Pearson correlation coefficient to calculate the correlation between various external factors and ride-hailing order data, and analyses the important factors affecting ride-hailing order volume through correlation analysis. In order to improve the stability of the network, a residual module is added. The results show that the model constructed in this paper has good prediction accuracy. The study shows the incorporation of multi-source data can effectively improve the prediction accuracy of the online ride-hailing prediction model.

Keywords: online ride-hailing demand; gated recurrent unit; GRU; convolutional neural network; CNN; travel demand.

DOI: 10.1504/IJSPM.2023.139774

International Journal of Simulation and Process Modelling, 2023 Vol.21 No.1, pp.22 - 32

Received: 29 Sep 2023
Accepted: 04 Jan 2024
Published online: 05 Jul 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article