In recent years, with the rapid development of national economy and transportation, the number of motor vehicles has increased dramatically, and the resulting traffic problems have aroused widespread concern. However, traffic flow is easily affected by weather, historical traffic flow and human factors, so traffic flow has the characteristics of randomness and nonlinearity, so using a single forecasting method to predict traffic flow can not meet the needs of modern intelligent transportation system. In order to improve the accuracy of traffic flow prediction, a traffic flow prediction model TF-VMD based on Variational Mode Decomposition (VMD) and Transformer is proposed in this paper. We use VMD method to decompose the traffic flow time series into k modes, and use Transformer model to predict the traffic flow in the next period of time. Finally, we superimpose the prediction results of each mode to obtain the final prediction results.Through experiments on real traffic flow data, compared with other prediction methods, the prediction results are improved in average absolute error, average absolute percentage error and equality coefficient, which effectively verifies the superiority of the model.
Traffic flow prediction model based on variational modal decomposition and Transformer
International Conference on Smart Transportation and City Engineering (STCE 2023) ; 2023 ; Chongqing, China
Proc. SPIE ; 13018
14.02.2024
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Europäisches Patentamt | 2023
|Modal Decomposition of Hamiltonian Variational Equations
British Library Conference Proceedings | 1994
|Multi-modal traffic flow prediction system based on AI
Europäisches Patentamt | 2024
|Traffic flow prediction method based on variational self-coding learning
Europäisches Patentamt | 2022
|