The invention discloses a traffic flow prediction method based on a multi-space-time diagram convolution network, and the method comprises the following steps: firstly, constructing a space-time assembly through employing ChebNet in combination with a gating cycle unit (GRU), so as to deeply mine the space-time correlation of nodes; secondly, respectively extracting sequence data of weekly correlation, day correlation and recent time, and inputting three space-time components to deeply mine time correlation among different time windows; and finally, fusing the space-time component and an encoder-decoder network structure (Encoder-Decoder) to form an MST-GCN model. A plurality of expressway data sets are used for an experiment, and the result shows that the performance of the new model is obviously better than that of a gating circulation unit model, a diffusion convolution circulation neural network (DCNN) model and a time-space diagram convolution network (T-GCN) model.
本发明公开的一种基于多时空图卷积网络的交通流预测方法,包括如下步骤:首先,利用切比雪夫图卷积(ChebNet)结合门控循环单元(GRU)构建时空组件以深度挖掘节点的时空相关性;其次,分别提取周相关、日相关、近期时间的序列数据,输入3个时空组件以深度挖掘不同时间窗口间的时间相关性;最后,将时空组件与编码器‑解码器网络结构(Encoder‑Decoder)融合组建MST‑GCN模型。利用多个高速公路数据集进行实验,结果表明新模型的性能明显优于门控循环单元模型、扩散卷积循环神经网络(DCRNN)模型和时空图卷积网络(T‑GCN)模型。
Traffic flow prediction method based on multi-space-time diagram convolutional network
一种基于多时空图卷积网络的交通流预测方法
2021-12-21
Patent
Elektronische Ressource
Chinesisch
Traffic flow prediction method based on multi-view space-time diagram convolutional network
Europäisches Patentamt | 2023
|Traffic flow prediction method based on space-time diagram convolutional network
Europäisches Patentamt | 2022
|Traffic flow prediction method based on space-time diagram convolutional network
Europäisches Patentamt | 2023
|Traffic flow prediction method of space-time diagram convolutional network
Europäisches Patentamt | 2024
|Traffic flow prediction method based on Transform space-time diagram convolutional network
Europäisches Patentamt | 2022
|