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

Seminar Announcement

Regression Splines: Applications and Fiducial Theory

Derek Sonderegger,Preliminary Ph.D., Department of Statistics, ColoradoState University

Friday, October 16, 2008

3:00 p.m. 223 Weber

ABSTRACT

Regression splines are a form of nonparametric function estimation is
used in many areas of applied statistics.  The selection of knot point
locations has not yet seen a satisfying theoretical conclusion,
perhaps because penalized regression splines provides an acceptable
workaround provided the knot locations are not of interest.  We first
introduce regression splines and provide an ecological situation where
the location of the knot point is of interest.  We then derive the
fiducial solution to regression spline problem and examine simulation
results of the method.  Finally, we derive the pointwise convergence
to normality of the parameter's fiducial distribution.

 

Advisory Committee:

Haonan Wang, Adviser

Jan Hannig, Co-Adviser

Hari K. Iyer, Committee Member

Barry R.Noon (Department of Fish, Wildlife & Conservation Biology)