|
Full Length Research Paper
|
|
A robust
method of estimating covariance matrix in multivariate data
analysis
G. M. Oyeyemi* and R. A.
Ipinyomi
Department of Statistics, University of Ilorin, Kwara State, Nigeria.
*Corresponding author. E-mail:
gmoyeyemi@yahoo.com,
ipinyomira@yahoo.co.uk.
Accepted 25 September, 2009
|
|
|
|
Abstract |
|
|
We proposed a robust method of estimating covariance matrix in
multivariate data set. The goal is to compare the proposed
method with the most widely used robust methods (Minimum Volume
Ellipsoid and Minimum Covariance Determinant) and the classical
method (MLE) in detection of outliers at different levels and
magnitude of outliers. The proposed robust method competes
favourably well with both MVE and MCD and performed better than
any of the two methods in detection of single or fewer outliers
especially for small sample size and when the magnitude of
outliers is relatively small.
Key words:
Covariance matrix, minimum volume ellipsoid (MVE), minimum
covariance determinant (MCD), mahalanobis distance, optimality
criteria.
|
