Paper ThAT7.3
Zhao, Xia (Beijing University of Civil Engineering and Architecture), Gao, yuan (Beijing University Of Civil Engineering And Architecture), Li, Zhihong (Beijing University of Civil Engineering and Architecture), Tang, Jiali (Traffic Monitoring Command Center,Nanchang), Zhao, Li (China Academy of Urban Planning)
IEF-BT-GCN: A High-Precision Rail Transit Passenger Flow Prediction Model with Complex External Factors
Scheduled for presentation during the Regular Session "Rail Traffic Management I" (ThAT7), Thursday, September 26, 2024,
11:10−11:30, Salon 15
2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC), September 24- 27, 2024, Edmonton, Canada
This information is tentative and subject to change. Compiled on October 8, 2024
|
|
Keywords Rail Traffic Management, Data Mining and Data Analysis, Other Theories, Applications, and Technologies
Abstract
With the continuous improvement of rail transit technology, the scale of urban rail transit network continues to expand, and the pressure on passenger flow operation and management is increasing. Under the comprehensive support of new technologies, new algorithms and new concepts, the research on passenger flow prediction methods of rail transit has received extensive attention. However, under the complex environment of urban rail transit passenger flow expansion and changing urban land layout, multi-granularity dynamic passenger flow prediction of rail transit "Network-Line-Station" is very difficult. At the same time, due to the interference of external factors, such as environmental factors, holidays, large-scale events, etc., the high degree of nonlinearity and uncertainty brings severe challenges to passenger flow forecasting. Therefore, it is urgent to construct a high-precision prediction model of rail transit with complex external factors.This paper proposes A Passenger flow forecasting model integrating complex external strong correlation factors (IEF-BT-GCN). The IEF-BT-GCN model was compared with the spatiotemporal graph convolutional model (BT-GCN) and some baseline models. Through comparison, it is found that the IEF-BT-GCN model has the following characteristics: (1) The IEF-BT-GCN model has high prediction accuracy for short-term and long-term prediction. (2) The prediction accuracy of the spatiotemporal environment enhancement feature model combined with multi-dimensional influencing factors is higher. (3) The IEF-BT-GCN model has higher prediction accuracy in peak prediction and low value prediction. (4) The IEF-BT-GCN model can better deal with emergencies. (5) The IEF-BT-GCN model is more suitable for predicting the situation with periodic changes
|
|