This paper proposes a network design method that uses the k-means clustering algorithm to determine the number of radial basis functions, function centers and widths of RBF neural networks. The output layer weights are determined by a set of linear equations, and the network parameters are optimized using the gradient descent method. In order to solve the problems of limited learning sample data, poor generalization ability of RBF networks and easy overfitting, a normalized network training method based on cross-validation is used in network training. Simulation experiments show that the generalization ability and classification accuracy of the network trained by this method are significantly improved, and the overfitting problem is effectively avoided.
You Might Like
Recommended ContentMore
Open source project More
Popular Components
Searched by Users
Just Take a LookMore
Trending Downloads
Trending ArticlesMore