Optimize CNN parameters based on grey wolf algorithm
Xianbin Ou · 2023
Recent studies have found that CNN can show quite good performance in various tasks such as recognition and classification, However CNN calculations require too many parameters, and the process of training is difficult. A large number of parameters may lead to over-fitting. In order to reduce the number of parameters and improve the accuracy of prediction, this Article proposed a model, on the basis of the traditional CNN algorithm, the gray wolf algorithm is used to automatically find optimal parameters, and the minimum error rate is used as the judgment criterion to find the optimal solution of the number of Convolutional layers and convolution kernels.