Math3034.03 Modern Regression Analysis
- Instructor: Peter Song.
- Office: N636 Ross; Phone: Ext. 33980;
- Email: email@example.com
- Time and Place Tue 12:30-2:30pm SLH E, Th 12:30-1:30pm S203 R
- Office Hours Thur 3-4pm or by appointment
- Problem Session TA: Rongcai Li
- Location: S110R, 12:30-1:30pm
- Text : Classical and Modern Regression with Applications by
R.M. Myers (Chapters 5 -- 9)
- Grading :
- 1 midterm 20%
- 4 assignments 5% each
- 2 data analysis projects 10% each
- final exam 40%
- Excuses :
No late homework will be accepted and no
makeup test will be given. If you cannot write a test for
medical reasons, you must provide a medical certificate and the weight
of the test will be transferred to the final exam. Otherwise you will
receive a zero for your test.
- Course Outline :
Continued to the previous course `Classical Regression Analysis',
this course will go on giving a further mathematical treatment of regression
analysis. Theory will be applied and demonstrated in data examples.
The first topic will center at the approach of model diagonsis including
anlysis of residuals (Chapter 6) and influence diagnostics (Chapter 7).
The second topic will focus on the development of regression models
that deal with some situations where nonstandard conditions or
violations of assumptions occur (Chapters 7 and 8). Finally, nonlinear
regression models (Chapter 9) will be introduced in this course if
Some assignments will involve the computer using Splus.
The introductory Splus guide: An Introduction to S and S-Plus by Phil
Spector, Duxbury Press, 1994, ISBN 0-534-19866-X, is on reserve in
Steacie Library and available in the bookstore.
All exams are closed book, except for the use of two
formula sheets, on which you may make whatever notes you wish.
You will need a calculator for all your work.
- York Card :
In this course, the computer lab S110 ROSS (GAUSS LAB) will be used.
You will need to use your York card to access the lab. The ATSG office
(N125 ROSS) can be contacted for door access problems.
- Data Analysis :
For the 2 data analysis porjects, students can choose data sets of
their own interest or use data sets provided by the instructor.
The purpose of these analyses is to develop data analysis skills and
apply the techniques taught to real data.
- Homework #1 Due February 4, 97.
- Problems 5.1, 5.4, 5.6, 5.8, 5.11, 5.15, 5.17.
- Homework #2 Due February 18, 97.
- Problems 6.1, 6.4, 6.6, 6.7. (The actual homework problems are
adapted from these exercises. For more details, see intructor Peter Song).
- Homework #3 Due
- Homework #4 Due
- Project #1 Due March 11, 97.
- This project is organised on the basis of individuals. Each of you is
required to write a project of data analysis with emphasis on applications of
the methodology of diagnostics developed in Chapters 5 and 6.
- You are encouraged to find data sets by yourselves in terms of your own
interests. However, if you are unable to do so, instructor will provide you
a data set for analysis. If you happen to get a "perfect" data set, meaning
no need for any improvement on issues such as outlier, high leverage points,
collinearity and transformation, please come to see the instructor.
- A project is primarily comprised of the following sections, Introduction,
Data, Preliminary analysis, Residual analysis, Final model and
- For more info, see the instructor.
- Project #2 Due
- Midterm will be held in class on March 4
- Final will be held on