OD-chebGCN: Traffic Flow Prediction Based on the Omni-Dimensional Dynamic Chebyshev Graph Convolution

In traffic flow prediction, in order to solve the problem of limited ability of traditional graph convolutional neural network (GCN) and two-dimensional convolutional operation to extract spatiotemporal correlation, a new traffic flow prediction model OD-chebGCN is proposed.Secondly, the ODconv component is used to dynamically pay attention to and adjust the convolution kernel, input space size, number of input channels and number of output channels to achieve effective extraction of time features.The spatio-temporal feature extraction module in this paper consists of the k-chebGCN component and the Odconv component, which collaborate and complement each other to complete the deep mining of the spatio-temporal correlation of traffic flow.In addition, the residual linkage between the modules helps to maintain the convergence and stability of the model.Finally, experiments are conducted on two public datasets.The results show that the prediction accuracy of OD-chebGCN is better than that of the baseline model, and it has an obvious competitive advantage in traffic flow prediction.