Title
Screening for detection of interactions in experiments with large numbers of factors
Speaker
Angela Dean, The Ohio State University
Abstract
Screening is the process of using designed experiments and statistical analyses to sift through a large number of factors in order to discover the few factors that influence a measured response (i.e. that are "active"). In this setting, conventional fractional factorial experiments typically require too many observations to be economically viable, especially when interactions are of interest. In this talk, strategies for screening in industrial experiments will be discussed, with the goal of identifying interactions with high probability. Web-based software will be described for the elicitation of information from subject specialists about which factors to include in the screening experiment as well as for the evaluation of different experimental designs. An example of group screening methodology will be illustrated through an experiment run at Jaguar Cars. Group screening in the computer experiments (deterministic) setting raises somewhat different issues and will be described briefly if time permits.
Meet the speaker in Room 212 Cockins Hall at 4:30 p.m. Refreshments will be served.
