Learning enhancement of radial basis function network with particle swarm optimization

Abstract

Back propagation (BP) algorithm is the most common technique in Artificial Neural Network (ANN) learning, and this includes Radial Basis Function Network. However, major disadvantages of BP are its convergence rate is relatively slow and always being trapped at the local minima. To overcome this problem, Particle Swarm Optimization (PSO) has been implemented to enhance ANN learning to increase the performance of network in terms of convergence rate and accuracy. In Back Propagation Radial Basis Function Network (BP-RBFN), there are many elements to be considered. These include the number of input nodes, hidden nodes, output nodes, learning rate, bias, minimum error and activation/transfer functions. These elements will affect the speed of RBF Network learning. In this study, Particle Swarm Optimization (PSO) is incorporated into RBF Network to enhance the learning performance of the network. Two algorithms have been developed on error optimization for Back Propagation of Radial Basis Function Network (BP-RBFN) and Particle Swarm Optimization of Radial Basis Function Network (PSO-RBFN) to seek and generate better network performance. The results show that PSO-RBFN give promising outputs with faster convergence rate and better classifications compared to BP-RBFN.