Skip to content

A collection of Matlab routines for constructing Radial Basis Function (Neural) Network models of NARX-type nonlinear dynamical systems from data. It also includes the possibility of incorporating prior information about the underlying system's steady states in the structure selection of the RBF models.

License

Notifications You must be signed in to change notification settings

maverickmath/RBF-NLD

Repository files navigation

RBF-NLD

Context: Steady-state performance constraints for dynamical models based on RBF networks. Engineering Applications of Artificial Intelligence, vol. 20, issue 7.

ctrsim: Builds a regression matrix with symmetric RBF centers; each line corresponds to a specific center; number of columns is equal to (nu>nul) + (ny>nyl).

estimaTeta: Estimates the parameters of an RBF model using classical least squares.

estimaTeta_r: Estimates the parameters of an RBF model subject to structural constraints.

igespacocp: Defines equally spaced RBF centers according to the model's input space.

kmeans: Distributes RBF centers over data class 'x' using classical k-means.

montaP: Builds a regression matrix.

montaPsim: Builds a "symmetrical" regression matrix.

mqermod: Outputs a symmetrical weight vector using constrained least squares.

mtrsres: Builds a matrix containing constraint values of both types (symmetrical weights and symmetrical centers).

myhouse: Implements the Error Reduction Rate criterion.

simulacao: Simulates a general RBF model.

simulacao_errn: Simulates an RBF model whose structure was selected according to the ERR criterion.

simulacao_r: Simulates a structurally constrained RBF model.

About

A collection of Matlab routines for constructing Radial Basis Function (Neural) Network models of NARX-type nonlinear dynamical systems from data. It also includes the possibility of incorporating prior information about the underlying system's steady states in the structure selection of the RBF models.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages