Study on Volterra filter based on group method of data handling and its application to flood forecast
Jinglin Qian · Journal of Zhejiang University(Engineering Science) · 2005
The delay time and embedding dimension were got by the method of auto-correlation function and the method of false neighbors respectively, then phase space was reconstructed and the maximal Lyapunov exponent was proved to be positive by calculation. The results showed that the flood series was chaotic,then (m-step)Volterra filter was used as the global predict function. The Volterra filter's coefficient exponentially increased with the number of the step, so it was more difficult to be applied to practical application. Group method of data handling(GMDH) was employed to model layer by layer and an ideal expression of Volterra filter based was got. The application of Shaxikou flood forecast showed that the Volterra filter based on GMDH was of high efficiency and of high accuracy.