Leonard A. Stefanski, PhD
Variable Selection and Measurement Error
Abstract: The talk will describe some creative uses of measurement error modeling concepts for the purpose motivating certain approaches to variable selection. The development from concept to method will be traced and methods so derived will be presented. Some previous work will be briefly reviewed and certain recent work more thoroughly described.
The talk will draw upon recent as yet unpublished work, and the published papers: Luo, Boos, and Stefanski (Technometrics 2006); Wu, Boos, and Stefanski (JASA 2007); Stefanski, Wu, and White (JASA 2014); and White, Stefanski, and Wu (JASA 2016); as well on.
Bio: Professor Stefanski received his Ph.D. in 1984 from the University of North Carolina, Chapel Hill. He was an assistant professor at Cornell prior to joining the faculty at NC State University in 1986, and a Visiting Professor in the Department of Biostatistics, School of Public Health, Harvard University (1988).
At NCSU he has served as the Graduate Program Co-Director, (2004-2007), Associate Department Head (2007-2011), and Interim Department Head (2016-2017). He is an elected Fellow of the American Statistical Association and served as Editor of Association’s flagship journal, The Journal of the American Statistical Association (2009-2012). He has served on a number of state and national committees, boards, and panels including the BEIR VII Committee on the Health Risks from Exposure to Low Levels of Ionizing Radiation, National Academy of Sciences (2001-2002), and the North Carolina Forensic Science Advisory Board (2012-2013). He has been awarded the D.D. Mason Department Award (2005) and the Cavell Brownie Mentoring Award (2013), a Drexel Professor of Statistics (2009), and the R. A. Fisher Distinguished Professor (2017).
Stefanski has coauthored three books, one edited volume, and numerous research arti- cles. His research, funded by the National Science Foundation and the National Institutes of Health, has been in the general area of statistical inference with particular emphasis on the analysis of data measured with error, robust statistical procedures, and variable and model selection. He has consulted extensively with environmental scientists at the US EPA (1981-2015).
Location and Address
University of Pittsburgh
1811 Wesley. Posvar Hall
230 South Bouquet Street
Pittsburgh, PA 15260