Applications to function approximation and system control based on wavelet networks
Ameni Jouila, Khaled Nouri · 2017
In the framework of nonlinear process modeling, we present conventional wavelet neural network used as nonlinear dynamic models type simple pendulum, this architecture is tuned by gradient-based techniques. Initialization procedure of wavelet coefficients will be established by the heuristic method. Performances of the initiated approach have been emphasized by results obtained from several experiences.