Optimizing EMG Classification through Metaheuristic Algorithms

This work proposes a metaheuristic-based approach to hyperparameter selection in multilayer perceptron classify EMG signals. The main goal of the study is improve performance model by optimizing four important hyperparameters: number neurons, learning rate, epochs, and training batches. proposed this shows that optimization using particle swarm gray wolf optimizer significantly improves classifying motion final achieves an average classification rate 93% for validation phase. results obtained are promising suggest may be helpful deep models other signal processing applications.