The invention discloses a traffic flow prediction method and system based on residual connection and space-time decomposition, and the method comprises the steps: carrying out the decoupling of the space-time dependence, respectively generating a space part prediction and a time part prediction, and finally carrying out the aggregation in an addition manner, and forming a final prediction result. False dependence possibly caused by the fact that unified hidden representation is adopted in the time dimension and the space dimension is avoided; on the other hand, layer-by-layer decomposition is introduced into a model architecture, and the flow data is decomposed into a plurality of predictable components layer by layer in the model, so that establishment of understanding of the data is facilitated, and future prediction is easier to make; and finally, the time identifier and the space identifier are introduced, so that the flow characteristic information is enriched, and the heterogeneity problem can be relieved.
本发明公开了一种基于残差连接与时空分解的交通流量预测方法和系统,它通过对时空依赖进行了解耦,分别产生空间部分预测和时间部分预测,最终以相加的方式聚合,形成最终的预测结果,避免了时间维度和空间维度采用统一隐藏表征可能导致的虚假依赖;另一方面,本发明将逐层分解引入到模型架构中来,流量数据在模型中逐层分解为若干个可预测的组件,有助于建立对数据的理解,从而更容易做出未来的预测;最后,本发明通过引入时间标识和空间标识,丰富了流量特征信息,有助于缓解异质性问题。
Traffic flow prediction method and system based on residual connection and space-time decomposition
一种基于残差连接与时空分解的交通流量预测方法和系统
2023-08-15
Patent
Elektronische Ressource
Chinesisch
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 / 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 flow prediction method based on residual decomposition recurrent neural network
Europäisches Patentamt | 2023
|Traffic flow prediction method of multi-network feature random fusion based on residual connection
Europäisches Patentamt | 2023
|Europäisches Patentamt | 2023
|Europäisches Patentamt | 2022
|Road network short-term traffic flow prediction method based on deep space-time residual network
Europäisches Patentamt | 2021
|