Rank-based partial autocorrelations are not asymptotically distribution-free

被引：0
|
作者：
Garel, B
Hallin, M
机构：
[1] Inst Natl Polytech, ENSEEIHT, Toulouse, France
[2] Free Univ Brussels, Dept Math, Brussels, Belgium
[3] Free Univ Brussels, ECARES, ISRO, Brussels, Belgium
来源：
STATISTICS & PROBABILITY LETTERS | 2000年 / 47卷 / 03期
关键词：
time series; autoregressive model; rank test; order identification; robust methods;
D O I：
10.1016/S0167-7152(99)00159-5
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：
Traditional AR order-identification methods rely on an inspection of partial correlograms. The rank-based version of partial correlograms is investigated here. Unlike the rank-based versions of Lagrange multiplier test statistics considered in Hallin and Purl (1994, J. Multivariate Anal. 50, 175-237), rank-based partial autocorrelations unfortunately are not asymptotically distribution-free, and thus cannot be used for order-identification purposes. (C) 2000 Elsevier Science B.V. All rights reserved.
引用
收藏
页码：219 / 227
页数：9
相关论文
共 50 条
12345