Building Personalized Transportation Model for Online Taxi-Hailing Demand Prediction

Zhiyuan Liu,Yang Liu,Cheng Lyu,Jieping Ye
DOI: https://doi.org/10.1109/tcyb.2020.3000929
IF: 11.8
2021-09-01
IEEE Transactions on Cybernetics
Abstract:The accurate prediction of online taxi-hailing demand is challenging but of significant value in the development of the intelligent transportation system. This article focuses on large-scale online taxi-hailing demand prediction and proposes a personalized demand prediction model. A model with two attention blocks is proposed to capture both spatial and temporal perspectives. We also explored the impact of network architecture on taxi-hailing demand prediction accuracy. The proposed method is universal in the sense that it is applicable to problems associated with large-scale spatiotemporal prediction. The experimental results on city-wide online taxi-hailing demand dataset demonstrate that the proposed personalized demand prediction model achieves superior prediction accuracy.
automation & control systems,computer science, cybernetics, artificial intelligence
What problem does this paper attempt to address?
### Problems the Paper Attempts to Solve This paper aims to address two key issues in large-scale online ride-hailing demand prediction: 1. **Capturing Personalized Spatiotemporal Information**: Existing literature lacks models that can effectively capture personalized spatiotemporal information for different regions. This paper proposes a model with two attention modules, designed to capture personalized information in both spatial and temporal dimensions. 2. **Impact of Network Architecture on Prediction Accuracy**: There is insufficient exploration in existing research on how network architecture affects the accuracy of ride-hailing demand prediction. This paper experimentally investigates the specific impact of different network architectures on prediction accuracy. ### Main Contributions 1. Proposes an online ride-hailing demand prediction system that includes two distinct attention modules to capture personalized spatial and temporal information for each region. 2. Explains the physical significance of the model and explores the impact of network architecture on the accuracy of ride-hailing demand prediction. 3. Combines traffic issues with the latest deep learning technologies, introducing innovative frameworks from computer vision and recommendation systems. ### Experimental Results Experiments conducted on a large-scale online ride-hailing dataset from Chengdu show that the proposed personalized demand prediction model performs excellently in terms of prediction accuracy.