Generalized radial basis function neural network based on an improved dynamic particle swarm optimization and AdaBoost algorithm

This paper proposes an improved dynamic particle swarm optimization algorithm, which uses a new and effective exponential decreasing inertia weight (EDIW) strategy. Based on the improved EDIW-PSO algorithm together with AdaBoost algorithm, we adjust the parameters (centers, widths, shape parameters and connection weights) of GRBF and present a novel hybrid EDIW-PSO-AdaBoost-GRBF model. Two application examples are given on the proposed model. The results obtained show that the proposed model is effective and feasible for prediction problems.