Regression Analysis (350-0-21)

Instructors

Feng Ruan

Meeting Info

University Hall 122: Tues, Thurs 12:30PM - 1:50PM

Overview of class

This is the Graduate level section of the 350 course.

This course is an applied statistics course, using the linear model to illustrate ideas in regressions.
The structure of the course is as follows: together we will work through linear models in increasing order of complexity to other nonlinear models, and we will develop techniques including model fitting, confidence intervals, hypothetical tests, model diagonstics, model selection etc along the way we shall discuss the ideas and concepts. Being an applied course, the focus of it is on solving real-life problems; along the development of tools we shall also talk about how to connect problems in practice with the established frameworks and methods developed in class, correctly and elegantly.

Registration Requirements

Prerequisite or co requisite: STAT 320-1

Learning Objectives

By the end of this class, the students are expected to know (1) the basic linear model and its glory: least squares, F-and T-tests, model diagnostics, (2) model selection: ridge regression, lasso, principal component regression, AIC/BIC, degrees of freedom, (3) M estimation and other forms of regression including robust regression and quantile regression

Evaluation Method

Homework assignments (30%), midterm exam (30%), and final exam (40%).

Class Materials (Required)

Plane Answers to Complex Questions: the Theory of Linear Models by Ronald Christensen
ISBN-10: 0387947671

Class Attributes

Formal Studies Distro Area

Enrollment Requirements

Enrollment Requirements: Prerequisites: STAT 201-0 or COMP_SCI 110-0 and STAT 202-0 or 210-0 or 232-0 or PSYCH 201-0 or IEMS 201-0 or 303-0. Co-requisite: STAT 320-1 or 383-0 or MATH 310-1 or 311-1 or 314-0 or 385-0 or ELEC_ENG 302-0 or IEMS 302-0.
Add Consent: Department Consent Required