Title :
On nonparametric kernel identification of nonlinear autoregression process
Author :
Kitaeva, A.V. ; Koshkin, G.M. ; Piven, I.G. ; Ryumkin, V.I.
Author_Institution :
Tomsk Polytech. Univ., Russia
fDate :
26 Jun-3 Jul 2001
Abstract :
In this paper, the piecewise-smoothed approximation for nonparametric regression estimation is proposed. The mean square convergence of this approximation from a dependent sample satisfying strong mixing conditions is proved. The main part of the asymptotic mean square error for the proposed modification of the kernel regression estimate is found. These results are used for the identification of the nonlinear autoregression process
Keywords :
autoregressive processes; convergence of numerical methods; mean square error methods; nonparametric statistics; piecewise linear techniques; asymptotic mean square error; dependent sample; kernel regression estimate modification; mean square convergence; nonlinear autoregression process; nonlinear autoregression process identification; nonparametric kernel identification; nonparametric regression estimation; piecewise-smoothed approximation; strong mixing condition; Convergence; Covariance matrix; Kernel; Mathematics; Mean square error methods; Random variables; State estimation; Statistics;
Conference_Titel :
Science and Technology, 2001. KORUS '01. Proceedings. The Fifth Russian-Korean International Symposium on
Conference_Location :
Tomsk
Print_ISBN :
0-7803-7008-2
DOI :
10.1109/KORUS.2001.975229
