Identification of linear and non linear curve fitting models using particle swarm optimization algorithm

摘要

Identification of approximate model of the physical systems can be achieved by fitting the data. In this paper particle swarm optimization algorithm (PSO) is used for linear and polynomial curve fittings. Data generated from the known models and curve fitting is done by PSO using reverse engineering mechanism at the initial stage. In this process of curve fitting, two types of inertia mechanisms are used in PSO for getting better results. Later, real time financial series forecasting is considered for validating the PSO estimated regression models. Results shows the dynamic inertia weight strategy based PSO yields better fitting and avoids additional decisions on control parameters.