## Wednesday July 27/10:30/Grande Ballroom

Invited Presentation 6

**Chair: James L. Phillips, Boeing Computer Services**
# Data-Adaptive Tuning of Extremely Large Dynamical System Models

In large scale models of dynamical systems such as occur in atmospheric or oceanic modeling, it is desirable to blend information from a dynamical model, from prior information concerning the behavior of the system being modeled, from known physical properties of the system, and from current or recent scattered, heterogenous, noisy and possibly sparse observational data. Generally, one is interested in the best possible estimate of the state of the system at some particular time, for general scientific purposes, or, the state of the system at the present time, for the purpose of forecasting. This may be as is the case for medium range global weather forecasting, by finding an approximate solution to a very large variational problem for a state or a state update.
The speaker will discuss the general problem of dynamically tuning such models by adjusting the various weights by cross validitory and related methods which are implementable in exremely large problems. (Based on joint work with D. R. Johnson, F. Gao, and J. Gong).

**Grace Wahba, Department of Statistics, University of Wisconsin, Madison**

Grace Wahba has a B.A. in Mathematics from Cornell where she had the good fortune to have Mark Kac for calculus; an M.A. in Mathematics from the University of Maryland, College Park, where Ryszard Syski was her thesis advisor; and a Ph.D. in Statistics from Stanford University where Emanuel Parzen was her thesis advisor. Professor Wahba worked for Operations Research, Inc. and IBM before she got her Ph.D., was a post doc at Stanford, and is a John Bascom Professor of Statistics at University of Wisconsin, Madison. She is the author of a book "Spline Models for Observational Data" (SIAM 1990) as well as author or co-author of about 100 articles, chapters and so forth, applied and theoretical. Professor Wahba thinks the most fun is when a real world problem raises interesting theoretical issues. She is very proud of her many high-achieving former students.