A TVAR parametric model based on WNN
Zhe Chen, Hongyu Wang, Qiu Tian-shuang · 2003
It is very difficult to describe a nonstationary random signal, to say nothing of processing it effectively. In recent years, the time-varying parametric model, especially, time-varying auto-regressive parametric model has been used widely. It is well known that a wavelet neural network has very good performance on function approximation. In this paper, the wavelet neural network is introduced into the time-varying auto-regressive parametric model, so a new time-varying auto-regressive parametric model based on wavelet neural network is presented. At the same time, a new algorithm for model parameters estimate is also presented. A few simulations indicate that the performance of the new time-varying auto-regressive parametric model is better than the old one.