Stat 550 - Regression, EO version originally recorded in Spring 2019

Instructor: Chris Williams, Room B - 16 Brink Hall

Mailing address: Department of Mathematics and Statistical Science, P.O. Box 441104, University of Idaho, Moscow ID 83844-1104

email: chrisw@uidaho.edu

Prerequisites: Math 330 (Linear Algebra), Stat 451 (Probability), Coreq of Stat 452 (Mathematical Statistics) or equivalent coursework.

Office Hours: MW 2:00-4:00 (in person or Zoom, the Zoom link is on the Canvas site) .

Webpage: http://www.webpages.uidaho.edu/~chrisw/stat550/

The webpage will contain announcements, summaries of lectures, lists of assignments with due dates, and other information.

Text: Applied Regression Analysis and Generalized Linear Models, third edition, by John Fox. A website for the text is at: http://socserv.socsci.mcmaster.ca/jfox/Books/Applied-Regression-3E/index.html . Among other items, online appendices on linear algebra, probability, and mathematical statistics for the text are available at:
http://socserv.socsci.mcmaster.ca/jfox/Books/Applied-Regression-3E/Appendices.pdf

Objectives: Theory and application of regression models including linear, nonlinear, and mixed models, and generalized linear models. Topics include model specification, point and interval estimators, exact and asymptotic sampling distributions, tests of general linear hypotheses, prediction, influence, multicollinearity, assessment of model fit, and model selection. 

Learning Outcomes: 
Understand the matrix-based formulation of the general linear model, including continuous and categorical predictors; 
Know the Gauss-Markov theorem; 
Ability to use matrix calculations to obtain the properties of least-squares estimators for the general linear model; 
Ability to use the geometrical representation of the general linear model to explain model properties; 
Ability to apply diagnostics for the general linear model and understand how to address problems revealed by diagnostics; 
Understand the nonlinear regression model and inference for this model;
Understand the generalized linear model and the use of likelihood-based approaches for inference with this model; 
Understand the linear mixed model with normal errors
Lecture Outline: We will cover chapters 5-15, 17, and 23 in the text, to be detailed in the lecture schedule.

We will use SAS and/or R throughout the course. 
Learning Outcomes will be assessed via homework sets, exams, and the project presentation.

Grading: Homework: 15%; Test 1 25%; Test 2 25%; Project 10%; Final 25%. Several homework sets will be assigned, but they will not all be handed in for a grade.

Academic Honesty: You should be aware policies of the University of Idaho concerning academic honesty (see Article II of the Student Code of Conduct). Breaches of academic honesty will not be tolerated and will result in an F for the course and referral to the Dean of Students for further disciplinary action. Student discussion on homework assignments is encouraged, but each student must write up their own solutions to be handed in, students are not to look at each other's written solutions. On exams, there is to be no communication between students, all questions and comments are to be addressed to me.

Center for Disability Access and Resources (CDAR) Reasonable Accommodations Statement: The CDAR coordinates services to meet the educational needs of students with temporary or permanent disabilities. Students needing accommodations to fully participate in a class should contact CDAR as soon as possible. All accommodations must be approved through CDAR prior to being implemented. To learn more about the accommodation process, visit CDAR at the Bruce M. Pitman Center Room 127, website at www.uidaho.edu/cdar or call 208-885-6307.