×

Nonparametric prediction for unbounded almost stationary processes. (English) Zbl 0768.62083

Nonparametric functional estimation and related topics, NATO ASI Ser., Ser. C 335, 389-403 (1991).
[For the entire collection see Zbl 0722.00032.]
A class of nonparametric predictors for unbounded mixing processes based on regression estimates, are studied. This class contains the classical convolution kernel and orthogonal series predictors. It is shown that it is not necessary to remove seasonal fluctuations before using the nonparametric prediction method. Finally, the author deals with the practical use of the method and compares it with the Box-Jenkins method.

MSC:

62M20 Inference from stochastic processes and prediction
62G07 Density estimation

Citations:

Zbl 0722.00032
PDFBibTeX XMLCite