SEMINARS
|
|
Fall 2007
STATISTICS
COLLOQUIUM
Wednesday, October 1, 2008
3:30-4:00—Refreshments
4:00-5:00—Talk
Yost Hall, Room 101
Curtis Tatsuoka, Ph.D.
Department of Quantitative Health Sciences, Cleveland Clinic
Sequential testing in classification on partially ordered models
Partially ordered sets are natural models for many statistical applications.
For instance, in modeling cognitive functioning, it may be natural to assume that certain states
have higher levels of functionality than others, and hence to assume that the respective states
follow a partial ordering. Another important application involves group testing, where it is of
interest to classify objects as defective or non-defective through the sequential testing of pooled objects.
A statistical framework will be described for the problem when there exists a “true” state among a collection
of states, and observations from sequentially selected experiments are used to identify it. An advantage
to sequentially selecting experiments is that testing burdens can be greatly reduced. Results on
asymptotically optimal sequential selection of experiments will be described in the context of Bayesian
classification problems when the parameter space is a finite lattice. Experiment selection rules and the
attainment of optimal rates of convergence will be discussed.
Finally, an analysis of neuropsychological testing data will be presented. Neuropsychological testing is used
to detect changes in cognition in clinical settings.
|
