A hybrid annual runoff prediction model using echo state network and gated recurrent unit based on sand cat swarm optimization with Markov chain error correction method

Abstract

ABSTRACT This study proposes a hybrid model based on the combination of Sand Cat Swarm Optimization (SCSO), Echo State Network (ESN), Gated Recurrent Unit (GRU), Least Squares Method (LSM), and Markov Chain (MC) to improve the accuracy of annual runoff prediction. Firstly, conduct correlation analysis on multi-factor data related to runoff to determine the input of the model. Secondly, the SCSO algorithm is used to optimize the parameters of the ESN and GRU models, and the SCSO-ESN and SCSO-GRU models are established. Next, the prediction results of these two models are coupled using LSM to obtain the preliminary prediction results of the SCSO-ESN-GRU model. Finally, the initial prediction results are corrected for errors using MC to get the final prediction results. Choose Changshui Station and Lanxi Station for experiments, and evaluate the predictive performance of the model through five evaluation indicators. The results show that the combined prediction model corrected by the MC achieved the optimal prediction performance at both experimental stations. This study emphasizes that using a combination prediction model based on Markov chain correction can significantly improve the accuracy of annual runoff prediction, providing a reliable basis for predicting annual runoff in complex watersheds.