Laplacian Deep Echo State Network Optimized by Genetic Algorithm

In recent years, the applications of Deep Echo State Network (DESN) are increasing in wind speed prediction, energy consumption prediction, temperature prediction, etc. However, when the reservoir structure of the DESN is too deep, its overall stability decreases, and its generalization ability decreases. For applications where the input dimension is much lower than the dimension of the reservoir structure, we proposed a novel strategy which uses genetic algorithms to optimize the Laplacian Echo State Network (LAESN) to solve the problems caused by the deep reservoir structure. We took the chickenpox case in New York City as a case, simulated the optimized DESN, and compared it with the three models-Echo State Network (ESN), DESN and LAESN, to verify its effectiveness.