A Hybrid Equilibrium Optimizer Algorithm and Its Optimization for LSTM

Aiming at the shortcomings of the equilibrium optimizer (EO) algorithm such as premature and slow convergence, a hybrid equilibrium optimizer (HEO) algorithm is proposed using a hybrid improvement strategy. Firstly, a chaotic reverse learning strategy is introduced to generate high-quality populations. Then, an elite Gaussian perturbation strategy is introduced to expand the exploration path and improve the optimization efficiency. Finally, a time-varying Levy perturbation strategy is introduced in the development phase to increase the ability of the algorithm to jump out of local optima. Simulation and comparative analysis of classical test functions show that the improved algorithm has better performance. The HEO is used for parameter optimization of the Long Short-Term Memory (LSTM) model. Meanwhile, in the classification of cardiovascular diseases, the HEO-LSTM is superior to other commonly models. It has the higher Classification accuracy and better model generalization performance.