The invention provides a heterogeneous graph traffic prediction method and system for strengthening road surrounding spatial features, and relates to the technical field of traffic prediction. The method comprises the following steps: extracting time characteristics of historical traffic information to obtain a first vector containing the time characteristics; extracting spatial features of the first vector to obtain a second vector containing the spatial features; the spatial features comprise road spatial features and road surrounding spatial information features; and performing convolution operation on the first vector and the second vector, and determining target traffic prediction information. According to the heterogeneous graph traffic prediction method and system for strengthening the spatial features around the road provided by the invention, traffic information can be accurately predicted.
本申请提供一种强化道路周边空间特征的异质图交通预测方法及系统,涉及交通预测技术领域。所述方法包括:对历史交通信息进行时间特征的提取,得到包含时间特征的第一向量;对所述第一向量进行空间特征的提取,得到包含空间特征的第二向量;所述空间特征包括道路空间特征和道路周边空间信息特征;对所述第一向量和所述第二向量进行卷积操作,确定目标交通预测信息。本申请提供的强化道路周边空间特征的异质图交通预测方法及系统,可以准确预测交通信息。
Heterogeneous map traffic prediction method and system for strengthening spatial features around road
强化道路周边空间特征的异质图交通预测方法及系统
2024-03-01
Patent
Electronic Resource
Chinese
European Patent Office | 2016
|ROAD TRAFFIC DEMAND PREDICTION DEVICE AND ROAD TRAFFIC DEMAND PREDICTION METHOD
European Patent Office | 2015
|Dynamic Spatial Correlation in Graph WaveNet for Road Traffic Prediction
Transportation Research Record | 2023
|