The invention discloses a traffic flow prediction method fusing trend and seasonality, and the method comprises the steps: 1) constructing a prediction network model based on a sequence-to-sequence framework, and the prediction network model comprises a trend prediction sub-model and a fusion prediction sub-model; 2) calculating a trend factor sequence and a seasonal factor sequence of a reference special period through a trend prediction sub-model by using data of previous years and data of a target special period lacking in the current year, and training a fusion prediction sub-model; and 3) calculating a trend factor sequence and a seasonal factor sequence of the target special period through the trend prediction sub-model, and inputting the trained fusion prediction sub-model to obtain traffic flow prediction data of the target special period. According to the method, trend and seasonal factors can be effectively fused under the condition that only single-year historical observation data exist, and the prediction accuracy and robustness are improved.
本发明公开了一种融合趋势与季节性的交通流量预测方法,包括步骤:1)构建基于序列到序列构架的预测网络模型,包含趋势预测子模型和融合预测子模型;2)利用往年数据和当年缺少目标特殊时期的数据,通过趋势预测子模型计算参考特殊时期的趋势因素序列与季节性因素序列,训练融合预测子模型;3)通过趋势预测子模型计算目标特殊时期的趋势因素序列与季节性因素序列,输入训练后的融合预测子模型,得到目标特殊时期的交通流量预测数据。本发明能够在仅具备单年度历史观测数据的条件下,有效融合趋势与季节性因素,提升预测的准确性与稳健性。
Trend and seasonal traffic flow prediction method
一种融合趋势与季节性的交通流量预测方法
2025-04-11
Patent
Electronic Resource
Chinese
A Model Based on Trend-Seasonal Decomposition and GCN for Traffic Flow Prediction
Springer Verlag | 2023
|European Patent Office | 2025
|Traffic flow trend prediction method, system and device and storage medium
European Patent Office | 2024
|Traffic flow prediction method based on time-space trend perception information fusion
European Patent Office | 2025
|Traffic flow prediction method and device based on trend similarity, medium and equipment
European Patent Office | 2024
|