"Everything should be made as simple as possible, but not simpler." - Albert Einstein

Seminar Announcement

Fiducial Inference Applications to a Maximum Mean and the Generalized Pareto

Damian Wandler, Ph.D. Candidate, Department of Statistics, Colorado State University .

Wednesday, March 11, 2009

9:00 am 223 Weber

ABSTRACT

Inference on the maximum mean of a multivariate normal is used frequently in testing for non-inferiority of a non-antiarrythmic drug when compared to a placebo in terms of the QT interval length. In this talk we will develop a new inference technique for the maximum mean based on an extension of R. A. Fisher's fiducial inference methods. We propose a fiducial based model for the maximum mean and conducted a simulation study to show that this method possesses promising empirical properties when compared to two competing methods. Furthermore, we proved that the fiducial based confidence intervals have asymptotically correct coverage. 

We will also speak to our future work that will be done with the generalized Pareto distribution. Including proving asymptotic properties for the fiducial model that we derived.

Advisory Committee:

Dr. Jan Hannig (Advisor)
Dr. Hari Iyer (Co-advisor)
Dr. Dan Cooley (Committee Member)
Dr. Edwin Chong (Electrical and Computer Engineering, Outside Member)