Robust Nonparametric Methods
for the General Linear Model
Professor Joseph W. McKean, Western Michigan University
Professor Jeff Terpstra, North Dakota State University
Abstract:
"Robust Nonparametric Methods for the General Linear Model", a two-day short course taught by Professor Joseph W. McKean of Western Michigan University and Professor Jeff Terpstra of North Dakota State University, will precede the conference on October 16 & 17. Robust rank-based procedures for linear models have their roots in traditional nonparametric methods for location problems. They offer the user a complete analysis of a linear model, including estimation, diagnostic checks for quality of fit, confidence intervals and regions, and tests of general linear hypotheses. This analysis is similar to the traditional least squares analysis, only unlike least squares, it is based on a fit that is robust to outliers in the Y-space. In this tutorial Profs. McKean and Terpstra discuss the rank-based analysis of the simple location problems, one- and two-sample problems, sample size determination, analysis for general linear models, covariance models, mixed models, and time series. The instructors will make use of the freeware R throughout the course.
Participants should visit http://www.ndsu.nodak.edu/instruct/terpstra/ACAS/
for important information and downloads related to this course.
The short course is a free service offered to conference registrants. No additional fees are required beyond conference registration.