Runoff Prediction Based on Deep Residual Shrinkage Long Short-term Memory Network

Abstract Aiming at the lack of sufficient meteorological data for runoff prediction in areas with few data, accuracy is improved by using soft thresholding under deep attention mechanism, and a residual shrinkage long-short memory network model constructed combined watershed characteristic data. The validated 134 watersheds spatial correlations CAMELS dataset. According to experimental findings, more accurate than LSTM model, HIV mHM predicting regions little an 86%.