Genetic learning of functional link networks

A genetic learning algorithm is proposed for supervised learning of Guassian-type of functional link networks (FLN), where Gaussian functions are used in the functional nodes. The parameters to be adjusted using genetic approach are weights between input layer and functional nodes, and parameters, i.e., center and width, of Gaussian functions (radial basis functions) in the functional nodes. Genetic coding is used for combining evolution of weights and Gaussian parameters in a single chromosome. Singular value decomposition (SVD) is used for computing the weights in the output layer. The proposed approach is efficient in terms of computational efficiency and time complexity as demonstrated with several benchmark datasets. The simulations indicate that proposed approach yields consistent results and near optimal solution, which is superior to previous approaches.