A New Strong Convective Precipitation Forecasting Method Based On Attention Mechanism And Spatio-temporal Reasoning

This article presents a novel forecasting method for strong convective precipitation using a combination of ResNet, attention mechanism, and Bidirectional Long Short-Term Memory. The method effectively extracts key information and captures the time-order relationship of sequence data, resulting in improved accuracy compared to existing models.