Adaptive Graph Convolution in Spatio-temporal Synchronous Networks for Traffic Flow Prediction

The field of traffic flow prediction is vital for smart city development and traffic management. However, accurately predicting spatio-temporal data is challenging due to its complexity and potential heterogeneity. To address this, a novel model called the spatio-temporal synchronous graph convolutional network (STASCN) has been introduced. This model effectively captures local spatio-temporal relationships and incorporates an adaptive graph learning approach to capture node-specific spatial and temporal correlations in a more focused manner. and also outperforms other methods in terms of prediction accuracy in real-life datasets.