|
Activity Number:
|
509
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 10, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Quality and Productivity
|
| Abstract - #305445 |
|
Title:
|
Goodness-of-Fit Testing and Pareto-Tail Estimation
|
|
Author(s):
|
Yuri Goegebeur*+ and Jan Beirlant and Tertius de Wet
|
|
Companies:
|
University of Southern Denmark and K.U. Leuven and University of Stellenbosch
|
|
Address:
|
JB Winslows Vej 9B, Odense, DK5000, Denmark
|
|
Keywords:
|
extreme value index ; quantile-quantile plot ; kernel statistic ; goodness-of-fit
|
|
Abstract:
|
In this contribution, a general kernel goodness-of-fit test statistic for assessing whether a sample is consistent with the Pareto-type model is introduced. The derivation of the class of statistics is based on the close link between the strict Pareto and the exponential distribution and puts some of the available goodness-of-fit procedures for the latter in a broader perspective. The limiting distribution for this general kernel statistic will be derived under mild regularity conditions and important special cases will be investigated in greater depth. The relation between goodness-of-fit testing and the optimal selection of the sample fraction for tail estimation (e.g., using Hill's estimator) is examined. The methodology is illustrated with a practical study.
|
