Intelligent transportation systems have been widely popularized in highway service areas, using optical sensors to collect real-time traffic flow data. It is crucial to conduct predictive analysis on the collected traffic flow data in order to further improve the service capacity and quality of highway service areas. Existing researches have not fully considered the section traffic survey data of highway near the entrance of service area, this paper comprehensively considers the section traffic survey data, section vehicle speed, and other important time-related features of the highway, and proposes the GPBO-LightGBM model for predicting the traffic survey data of entering the service area. It is proved experimentally that the predict performance of GPBO-LightGBM model are improved by 65.8% and 14.6% compared with XGBoost and LightGBM, respectively. In addition, through the feature correlation analysis, the traffic flow survey data of the highway section, the vehicle speed of the section and the traffic flow survey data entering the highway service area have significant positive correlation.
Trend prediction of traffic flow survey data in intelligent service areas based on GPBO-LightGBM
Second International Conference on Big Data, Computational Intelligence, and Applications (BDCIA 2024) ; 2024 ; Huanggang, China
Proc. SPIE ; 13550
2025-03-20
Conference paper
Electronic Resource
English
Short-time traffic flow prediction based on a combined GAT-VMD-LightGBM prediction model
British Library Conference Proceedings | 2023
|Flow Prediction of Aero Engine Driven Pump Based on QPSO-LightGBM
British Library Conference Proceedings | 2022
|Flight delay prediction based on LightGBM
IEEE | 2021
|Flow Prediction of Aero Engine Driven Pump Based on QPSO-LightGBM
Springer Verlag | 2021
|