These Aren't Your Mothers and Fathers Experiments
Dr. Thomas J. Santner
The Ohio State University
The Department of Mathematics at the University of Dayton presents
The 10th Annual Kenneth C. Schraut Memorial Lecture
Saturday, November 7, 2009
|
|
These Aren't Your Mothers and Fathers Experiments Dr. Thomas J. Santner The Ohio State University
|
Abstract: Informal experimentation is as old
as humankind. Statisticians became seriously involved in the conduct of
experiments during the early 1900s when they devised methods for the design of
efficient field trials to improve agricultural yields. During the 1900s
statistical methodology was developed for many complicated sampling settings and
a wide variety of design objectives. For example, Taguchi emphasized designing
experiments for the goal of creating "robust product designs," i.e., products
whose performance was as impervious as possible to variability in the
constituent components of the final product. Within the last 15 years
experimentation has been heavily influenced by high speed computing and
developments in numerical algorithms. The starting place for this computational
influence is the work of applied mathematicians who have created sophisticated
theoretical models of the input-output relationships of many engineering and
biological systems. The implementations of such models in computer codes are
routinely used as surrogates (or adjuncts) for physical experiments. This use of
a computer code is often termed a "computer experiment." Automobiles, airplanes,
and prosthetic devices are but a few of the many products whose development
relies heavily on computer experiments. This talk will describe the breadth of
applications of computer experiments and sketch the framework used to think
about the analysis of data from them. We will also give some (unsolicited)
advice about graduate school and life beyond the undergraduate degree.
About the speaker:
Professor Santner, who received his B.S. (1969) from the University of
Dayton and his M.S. (1971) and Ph.D. (1973) from Purdue University is a Professor
in the Department of Statistics at The Ohio State University. Professor Santner was a
member of the faculty at Cornell University from 1973-1989 before joining the
faculty at the Ohio State University in 1990. He served as chair of the
Department of Statistics during 1992-2000 and he is also a former Director of
the Department's Statistical Consulting Service. He currently serves as
chair of the Department of Biostatistics in the Ohio State University College of
Public Health. He has held various visiting
positions including a position as a Fulbright Scholar at Ludwig Maximilians
Universität in Munich, Germany, and visiting positions at Cornell University and
Duke University.
Professor Santner's research interests are in the design of experiments and the analysis of discrete data. He is the co-author of three books, The Statistical Analysis of Discrete Data (with Diane Duffy), Springer-Verlag, Inc. (1989), Design and Analysis of Experiments for Statistical Selection, Screening, and Multiple Comparisons (with R. E. Bechhofer and D. M. Goldsman), J. Wiley and Sons (1995), and The Design and Analysis of Computer Experiments (with Brian Williams and Bill Notz), Springer-Verlag, Inc. (2003).