AGA-GRU: An Optimized GRU Neural Network Model Based on Adaptive Genetic Algorithm

Abstract

The weight adjustment of gated recurrent unit (GRU) network depends on the gradient descent algorithm heavily, therefore this paper proposes an improved GRU neural network model based on adaptive genetic algorithm (AGA-GRU) to solve this problem. In this model, AGA is used to construct the optimization system, and the parameters of neural network model are optimized to improve the prediction performance. The results on UCI dataset show that the prediction accuracy of AGA-GRU model is significantly improved, and the generalization performance is stronger.