IASC - A New Robust 2T Monitoring Chart Based On The Robust Bootstrapped Singular Value Decomposition
| Date | 28 Jun 2024 |
| Time | 14:00 GMT+02:00 - 15:00 GMT+02:00 |
| Level of instruction | Intermediate |
| Instructor |
Prof. Emmanuel John Ekpenyong
Dr. Chisimkwuo John
|
| Registration fee | |
The Hoteling’s 2T chart regarded as the classical multivariate charting technique has shown to be of great usefulness in industrial multivariate monitoring. Its usage in the process industry becomes apparent as many modern industrial processes are highly multivariate in nature. Monitoring products in order to maintain quality employs the 2T monitoring charts because of the flexibilities in using the chart. However, the multivariate term in the Hoteling’s 2T requires the validation of the normality assumption and deviation from multivariate normality maybe a limitation due to multivariate outliers, correlation, incomplete or short data runs, but to mention a few. Meanwhile, many manufacturing systems struggle with meeting the goals of high turnout of products as well as reducing downtime in the system, making these setbacks inevitable. In this study, a new multivariate 2T chart that is based upon the robust bootstrapped Singular Value Decomposition (SVD) system has been introduced to ameliorate these setbacks by cropping out a new robust 2T chart termed RobBootSVD 2T . The new 2T is appraised with the existing 2T and robust 2T charts that are based upon the ordinary SVD and the robust SVD respectively. In the overall, the new charting system has been appraised using both real dataset and various contaminated datasets and the results obtained shows that the new RobBootSVD 2T outperforms the ordinary 2T and is comparable to the existing robust SVD 2T.
Instructors
About the instructor
Prof. Emmanuel John Ekpenyong is a Professor of Sampling Theory and Survey Methods in the Department of Statistics, Michael Okpara University of Agriculture, Umudike, Nigeria. He is a member of the following professional organizations among others: CISON, ISI, IASC, IBS, IAOS. He is a former member of the governing council of Nigerian Statistical Association. His further research interests are on Time series analysis and data science.
About the instructor
Dr. Chisimkwuo John, [BSc. (Umudike), BComHons, MCom (Stellenbosch) and Ph.D (Umudike)] is a Senior Lecturer in the Department of Statistics, Michael Okpara University of Agriculture, Umudike (MOUAU) since 2013. He is a major in the area of Applied Industrial Statistics and Data Analytics where he obtained his Ph.D. He has taught both postgraduate and undergraduate modules like the Multivariate Statistical Analysis, Statistical Computing using R/Python, Statistical Quality Control, Design of Experiments, and some other service Statistics courses for several years.
Dr. John was the pioneer Head of the MOUAU Statistics Laboratory where he helps postgraduate students to program and implement new findings and also teaches advanced procedures using the R programming software and other related softwares like Python, SAS, STATISTICA, SQL, Excel, etc. He is also currently the postgraduate coordinator in the Department of Statistics, MOUAU and has supervised several undergraduate students and a Masters student in Applied Statistics, Machine learning and Data Science areas. Apart from making good contributions in paper presentations / publications, Dr. John has also provided various statistical consultations and workshop trainings to many institutions and multinationals like the National Insurance Corporation of Nigeria (NICON), Grambling State University, LA, USA, Selected Postgraduate Statistics Group from Nnamdi Azikiwe University, Awka, UNFPA/Bureau of Statistics in Abia State, Nigerian Statistical Association (NSA), Laboratory for Interdisciplinary Statistical Analysis (LISA), and the British American Tobacco (BAT), but to mention a few.