Statistics and Risk Management Seminar - 19/6/2014

Thursday, 19 June 2014, 2:00 p.m.

Lecturer: Senem Şahan Vahaplar, Assistant Professor, Dokuz Eylül University, Turquia

Title: "An Application on an Order Statistics Based Distribution-Free Control Charts"

Local: Sala de Seminários - Edifício VII
Faculdade de Ciências e Tecnologia, Quinta da Torre, Caparica

Abstract: In statistical process control, causes of variation in the processes can be detected by using control charts. The control limits of traditional Shewhart control charts are based on the assumption that the quality characteristic is distributed normally. But frequently, this normality assumption is not satisfied in real life data and this affects the statistical properties of standard control charts.

When this is the situation, that is when normality assumption is not satisfied, distribution-free or non-parametric control charts can be used. Distribution-free control charts find wide applications in statistical process control and there are several papers published in literature. Using distribution-free control charts provide many advantages to the researchers. The first of these advantages is that, they do not need to assume a particular probability distribution. Besides, in-control run length distribution is the same for all continuous distributions. Also, they are more robust and more efficient in detecting changes when the underlying distribution is non-normal.