|
Activity Number:
|
289
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #306838 |
|
Title:
|
A Markov Chain Monte Carlo Approach for Finding the Minimum Volume Ellipsoid
|
|
Author(s):
|
J. Brian Gray*+ and Eric B. Howington
|
|
Companies:
|
The University of Alabama and Coastal Carolina University
|
|
Address:
|
Department of ISM, Tuscaloosa, AL, 35487-0226,
|
|
Keywords:
|
data mining ; outlier ; minimum covariance determinant ; robust diagnostics ; robust covariance estimation
|
|
Abstract:
|
The minimum volume ellipsoid (MVE) is a useful construct for computing robust multivariate outlier diagnostics and robust covariance matrix estimates. Exact computation of the MVE is impractical for all but the smallest data sets. A variety of approaches have been developed for approximating the MVE. In this presentation, we introduce a new MVE approximation method based on a Markov Chain Monte Carlo approach. This method can also be used to find the minimum covariance determinant (MCD) estimator. The proposed method is compared to existing techniques using real and simulated data.
|
