The invention provides a traffic prediction method based on a sub-hypergraph and self-distillation. The method comprises the following steps: firstly, clustering the whole data according to different traffic characteristics to construct a sub-hypergraph; then, constructing a teacher network by using the sub-hypergraphs, and extracting different attribute features contained in the data; meanwhile, a student network is constructed based on overall data and a traffic road topological structure, and global structural features of a traffic road network are learned. And finally, migrating the data features learned by the teacher network to the training process of the student network by using a self-distillation method to obtain a final prediction result. According to the invention, evaluation is carried out on several real-world traffic data sets. Experimental results prove that the method is effective, and prediction errors are reduced.
本发明提出了一种基于子超图和自蒸馏的交通预测方法。本发明首先根据不同的交通特征对整个数据进行聚类来构建子超图。随后,利用这些子超图构建教师网络,提取数据中蕴含的不同属性特征。同时,基于整体数据和交通道路拓扑结构构建学生网络,学习交通路网的全局结构特征。最后,利用自蒸馏方法将教师网络学习到的数据特征迁移到学生网络的训练过程中,得到最终的预测结果。本发明在几个真实世界交通数据集上进行了评估。实验结果证明了本发明的有效性,减小了预测误差。
Traffic prediction method based on sub-hypergraph and self-distillation
一种基于子超图和自蒸馏的交通预测方法
2024-10-08
Patent
Electronic Resource
Chinese
IPC: | G06Q Datenverarbeitungssysteme oder -verfahren, besonders angepasst an verwaltungstechnische, geschäftliche, finanzielle oder betriebswirtschaftliche Zwecke, sowie an geschäftsbezogene Überwachungs- oder Voraussagezwecke , DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES / G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS |
Traffic prediction method and system based on adaptive hypergraph
European Patent Office | 2023
|Equivalent hypergraph construction method for traffic flow prediction
European Patent Office | 2025
|Deep learning traffic flow prediction method based on time perception hypergraph
European Patent Office | 2024
|Traffic flow prediction method based on bidirectional GRU hypergraph convolution model
European Patent Office | 2024
|