|
Activity Number:
|
339
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Risk Analysis
|
| Abstract - #308751 |
|
Title:
|
A New Estimator for the Extremal Index Based on Scaling and Resampling
|
|
Author(s):
|
Kam Hamidieh*+ and Stilian Stoev and George Michailidis
|
|
Companies:
|
University of Michigan and University of Michigan and The University of Michigan
|
|
Address:
|
528 S State, Ann Arbor, MI, 48104,
|
|
Keywords:
|
Extremal Index ; Clusters ; Heavy Tails ; Extreme Value Theory
|
|
Abstract:
|
The extremal index is the key parameter quantifying the clustering of the extremes of a stationary sequence. In this talk, we propose a new estimator of the extremal index based on the scaling properties of the block maxima of heavy tailed data. We discuss some of its properties and demonstrate its competitive nature through an extensive simulation study. An application that explores the structure of a financial data set is also presented. This is joint work with Stilian Stoev and George Michailidis.
|
