Title:
PROPOSED HOTELLING’S T2 CONTROL CHARTS BASED ON LOCATION AND SCATTER MATRIX.

Authors:
*Obafemi, Oluwafemi Samuel, Oyeyemi, Gafar Matanmi and Bolarinwa, Folashade Adeola

Abstract:
In many quality control settings, the product (process) under examination may have two or more correlated quality characteristics; hence, an appropriate approach is needed to simultaneously monitor all the quality characteristics. The Hotelling T 2 control chart based on the usual sample mean vector and variance - covariance matrix performs poorly, especially when there are multiple out of control points in the multivariate data set. Several alternative methods have been proposed, this includes methods based on the minimum volume ellipsoid (MVE) and the minimum covariance determinant (MCD) among many other methods. These control charts are powerful in detecting a reasonable number of outlying data.in this paper we propose a modified Hotelling T2 control charts using the eigen-values obtained from scatter matrix/ variancecovariance matrix of the multivariate data. The methods were used on a real-life data set. The studies show that this method outperforms the classical Hotelling T2 control charts and compete well with charts based on MVE and MCD for a small number of observations when the number of out of control points was increased

Keywords:
Hotelling T2 , Control charts, Outliers, eigen-value, quality Characteristics, Mean shift.