Sankhya: The Indian Journal of Statistics

1994, Volume 56, Series A, Pt. 3 , pp. 500-511

IDENTIFYING THE LOCATION SHIFT OUTLIERS THAT MATTER

By

M. A. ALI, University of Bahrain

SUMMARY. In this article the effect of the outliers and the effect of deleting observations on the least squares estimate of the vector of unknown regression coefficients are assessed. Location shift outliers in a linear model are classified into two categories-moderate location shift outliers and the extreme location shift outliers. The impact of the moderate and extreme location shift outliers on Cook's (1977) distance is examined. Identification of the extreme location shift outliers is discussed. A simple example explaining our procedure is given.

AMS (1980) subject classification. 62J05.

Key words and phrases. Design matrix, mean squared error, non-central $F$ distribution, Cook's distance.

Full Paper (PDF).

This paper in Mathematical Reviews.