ABSTRACT Time‐evolving traffic flow forecasting is playing a vital role in intelligent transportation systems and smart cities. However, the dynamic traffic flow forecasting is a highly nonlinear problem with complex temporal‐spatial dependencies. Although the existing methods have provided great contributions to mine the temporal‐spatial patterns in the complex traffic networks, they fail to encode the globally temporal‐spatial patterns and are prone to overfitting on the pre‐defined geographical correlations, and thus hinder the model's robustness in the complex traffic environment. To tackle this issue, in this work, we proposed TSFusion , a multi‐grained temporal‐spatial graph learning framework to adaptively augment the globally temporal‐spatial patterns obtained from a crafted graph transformer encoder with the local patterns from the graph convolution by a crafted gated fusion unit with residual connection techniques. Under these circumstances, our proposed model can mine the hidden global temporal‐spatial relations between each monitor station and balance the relative importance of local and global temporal‐spatial patterns. Experiment results demonstrate the strong representation capability of our proposed method, and our model consistently outperforms (more than 11.5% for MAE) other strong baselines on various real‐world traffic networks.
Lin et al. (Thu,) studied this question.