ON RESIDUAL SUMS OF SQUARES IN NONPARAMETRIC AUTOREGRESSION

被引：14

作者：

CHENG, B ^{[1
]}

TONG, H ^{[1
]}

机构：

[1] UNIV KENT,INST MATH & STAT,CORNWALLIS BLDG,CANTERBURY CT2 7NF,KENT,ENGLAND

来源：

STOCHASTIC PROCESSES AND THEIR APPLICATIONS | 1993年 / 48卷 / 01期

关键词：

BIAS; CROSS VALIDATION; KERNEL; NONPARAMETRIC AUTOREGRESSION; RESIDUAL SUM OF SQUARES; U-STATISTICS;

D O I：

10.1016/0304-4149(93)90112-H

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

By relying on the theory of U-statistics of dependent data, we have given a detailed analysis of the residual sum of squares, RSS, after fitting a nonlinear autoregression using the kernel method. The asymptotic bias of the RSS as an estimator of the noise variance is evaluated up to and including the first order term. A similar quantity, the cross validated residual sum of squares obtained by 'leaving one out' in the fitting is similarly analysed. An asymptotic positive bias is obtained.

引用

页码：157 / 174

页数：18

共 22 条

[1] STATISTICAL PREDICTOR IDENTIFICATION
AKAIKE, H
[J]. ANNALS OF THE INSTITUTE OF STATISTICAL MATHEMATICS, 1970, 22 (02) : 203 - &
[2] IDENTIFICATION OF NONLINEAR TIME-SERIES - 1ST ORDER CHARACTERIZATION AND ORDER DETERMINATION
AUESTAD, B
TJOSTHEIM, D
[J]. BIOMETRIKA, 1990, 77 (04) : 669 - 687
[3] Bradley R. C., 1986, DEPENDENCE PROBABILI
[4] CHENG B, 1992, J ROY STAT SOC B MET, V54, P427
[5] CHENG B, IN PRESS BIRTHDAY VO
[6] ON U-STATISTICS AND V MISES STATISTICS FOR WEAKLY DEPENDENT PROCESSES
DENKER, M
KELLER, G
[J]. ZEITSCHRIFT FUR WAHRSCHEINLICHKEITSTHEORIE UND VERWANDTE GEBIETE, 1983, 64 (04): : 505 - 522
[7] GJORFI L, 1990, LECTURE NOTES STATIS
[8] Hannan E. J., 1986, J TIME SER ANAL, V7, P27
[9] HOW FAR ARE AUTOMATICALLY CHOSEN REGRESSION SMOOTHING PARAMETERS FROM THEIR OPTIMUM
HARDLE, W
HALL, P
MARRON, JS
[J]. JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1988, 83 (401) : 86 - 95
[10] HARDLE W, 1985, ANN STAT, V13, P1461

← 1 2 3 →