Department of Statistics Seminar
North Carolina State University
presents
Dr. Geoffrey Vining
Virginia Tech
"An Overview on the Current Work on Industrial Split-Plot Experiments"
ABSTRACT
Many industrial processes involve two types of factors:
hard-to-change and easy-to-change, which lead to split-plot structures.
This structure results in two different error terms: one at the "whole
plot" level and another at the "subplot" level. Split-plot experiments are
very common in agriculture. However, agricultural and industrial
experiments have distinct but subtle differences. Agricultural experiments
tend to focus upon the estimation of contrasts while industrial experiments
often focus on the estimation of a response surface.
Of special interest to this talk is the estimation and testing associated
with fitting a second-order response surface model. Some of my recent work
looks at designs for which the ordinary least squares estimates of the model
are equivalent to the generalized least squares estimates, which turns out
to lead to interesting results.
Friday, April, 7, 2006
3:35 - 4:35 pm
206 Cox Hall
Refreshments will be served on the second floor of Dabney Hall (left of Room 222) at 3:00 pm.