Multifeature fusion traffic flow prediction based on deep learning

In order to accurately predict highway traffic flow under external meteorological disturbance, a multi feature fusion traffic flow prediction model based on deep learning algorithm is proposed. Firstly, the attention-embedded gated recurrent unit is used to capture the key time-dependent features of the traffic flow series, and the graph convolutional network is used to learn the spatial features of the traffic flow data. Secondly, The feature filtering mechanism and multi-layer fully connected network are used to extract external meteorological features, Finally, the fusion layer is defined to aggregate the time-dependent features, spatial features of traffic flow and external meteorological features, so as to realize highway traffic flow prediction. The experimental results based on the PeMS08 dataset show that compared with the baseline model, the RMSE of the proposed model is decreased by at least 2% in the single-step forecasting experiment. At the same time, it further proves the superiority and stability of the model in the multi-step prediction scenario.