Outliers Detection in Linear Regression Using Eigenvalues


Speaker: Kenan Matawie

Affiliation: University of Western Sydney

Time: Monday 21/11/2011 from 14:00 to 15:00

Venue: Access Grid UWS. Presented from Parramatta (EB.1.32), accessible from Campbelltown (26.1.50) and Penrith (Y239).

Abstract: Single and multiple outliers detection in linear regression are briefly reviewed and linked to this talk, which will present a new approach for detection using eigenstructure of the influential matrix that is defined as a normalised matrix of the differences between least squares and robust fit using Cook statistics. In this talk I will also present a developed outlier detection criteria and discuss it through.

Biography: Kenan Matawie is currently a Senior Lecturer in the School of Computing, Engineering and Mathematics, UWS. He is an academic statistician with strong interest in application of statistics, especially research in statistical modelling and advance data analysis and developing statistical methods motivated by real-world problems. Kenan is currently connected and contributing to two UWS research groups/centres and three national and international research centres, especially Erasmus Medical Centre, Department of biostatistics.