Advances and Applications in Statistics
Volume 19, Issue 2, Pages 97 - 111
(December 2010)
|
|
A GRAPHICAL QUALITY CONTROL PROCEDURE USING DATA DEPTH
Hajlaoui Mekki
|
Abstract: Quality is a multivariate concept by nature. So since the Mahalanobis work, the statisticians tried to develop devices that help in monitoring the quality process. Their efforts come to breed multiple procedures but all of them are dependent to multiple conditions to be applied in industry. Using the centre-outward ranking induced by the notion of data depth, we use a graphical method named DD-diagram to see the outlyingness of a given multivariate quality data. This one-dimensional curve in the plane is used to monitor a multivariate quality process. Hence, shifts in mean and/or in scale are analyzed through graphical representation of an observed sample. To see the benefits of this graphical method test, a real example from industry is presented. In addition, we suggest to rank data depths before sketching out the DD-diagram. |
Keywords and phrases: multivariate data, quality control, data depth, DD-diagram. |
|
Number of Downloads: 355 | Number of Views: 1059 |
|