A simple robust estimation method for the thickness of heavy tails

被引：45

作者：

Meerschaert, MM

Scheffler, HP

机构：

[1] Univ Dortmund, Fachbereich Math, D-44221 Dortmund, Germany

[2] Univ Nevada, Dept Math, Reno, NV 89557 USA

来源：

JOURNAL OF STATISTICAL PLANNING AND INFERENCE | 1998年 / 71卷 / 1-2期

关键词：

parameter estimation; heavy tails; regular variation; domains of attraction;

D O I：

10.1016/S0378-3758(98)00093-7

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：

We present a simple general method for estimating the thickness of heavy tails based on the asymptotics of the sum. The method works for dependent data, and only requires that the centered and normalized partial sums are stochastically compact. For data in the domain of attraction of a stable law our estimator is asymptotically log stable, consistent and asymptotically unbiased, and converges in the mean-square sense to the index of regular variation. (C) 1998 Elsevier Science B.V. All rights reserved.

引用

页码：19 / 34

页数：16

共 34 条

[1]

Anderson PL, 1997, ANN STAT, V25, P771

[2]

Bingham N. H., 1989, Encyclopedia of Mathematics and Its Applications, V27

[3]

Brockwell P.J., 1991, TIME SERIES THEORY M

[4] METHOD FOR SIMULATING STABLE RANDOM-VARIABLES [J].