Output details

Non-linear system identification using particle swarm optimisation tuned radial basis function models

Significance of output:

Here we propose a novel particle swarm optimizer for the parsimonious modelling. The approach is shown to be quite generic in that it can be applied to regression, classification and probability density function determination. The approach can be viewed as combining both the non-linear and linear learning methods and provides the greater modelling capability of the non-linear approach while offering the computational simplicity of the linear fixed-node approach. This offers significant advantages in terms of generalisation, mean square error, etc., over other known modelling approaches.