Syllabus.  
Statistics 201b:  Regression Analysis.<br>
Prof. Rick Paik Schoenberg. <br>

Lectures:  MW 3:00-4:20pm, MS 5203.<br><br>

Office hours:  Mondays, 1:00-2:20pm, MS 8965. Or by email appointment.<br><br>

email:  frederic@stat.ucla.edu<br><br>

Course webpage:  <a href="http://www.stat.ucla.edu/~frederic/201b/W07">
http://www.stat.ucla.edu/~frederic/201b/W07 </a><br><br>

Text: "Data Analysis and Graphics Using R" by John Maindonald and John Braun.

Description:  Applied regression analysis, with emphasis on general 
linear model (e.g., multiple regression) and generalized linear model
 (e.g., logistic regression). Special attention to modern extensions 
of regression, including regression diagnostics, 
graphical procedures, and bootstrapping for statistical inference.<br><br>

  
Grading: <br>
Class participation (5%).<br>
Homework (20%).<br>
Written Project (25%). <br>
Midterm. (20%).<br>
Final exam. (30%). <br>

The midterm is Monday, Feb 26, 2007, 3:00-4:20pm (in class). 
The final exam is Wednesday, March 21, 2007, 3:00pm-6:00pm.
Project due dates will be announced in class. 
Homework will be due each Wednesday, beginning with HW1 which is due
Wednesday Jan 17.<br><br>

Homeworks must be handed in at the beginning of class, or may be slipped under 
my office door (MS 8965) any time before class. Each homework assignment 
is graded out of 10 points. 
Homeworks handed in between 3:05 and 3:10 will be given a one-point deduction. 
Those handed in between 3:10 and 3:20 will be given a two-point deduction. 
Homeworks handed in between 3:20 and 4:20 will be given a three-point deduction.
Homeworks submitted after 4:20pm will not be accepted. Homeworks and projects 
must be submitted in hard copy, rather than by email or fax. <br><br>

Rough Outline:<br>
Week 1: Introductory material. Chapters 1-2.<br>
Week 2: Univariate modeling and inference. Chapters 3-4.<br>
Week 3: Regression with one predictor. Chapter 5.<br>
Week 4: Multiple linear regression. Chapter 6.<br>
Week 5: Linear model. Chapter 7.<br>
Week 6: Logistic regression and generalized linear models. Chapter 8.<br>
Week 7: Review and midterm.<br>
Week 8: Multi-level models, time series, and repeated measures. Chapter 9.<br>
Week 9: Tree-based classification and regression. Chapter 10.<br>
Week 10: Review.<br><br>

Homework 1 (due Wednesday Jan 17): 
1st edition: p26 #2, p27 #4, p50 #1, p51 #2, p51 #3, p51 #5.<br>
Homework 2 (due Wednesday Jan 24): 
1st edition: p103 #2, p103 #3, p104 #6, p105 #11, p132 #1, p132 #2<br>
Homework 3 (due Wednesday Feb 7): 
2nd edition: p243 #2, p244 #7, p244 #8, p245 #9, 
p245 #14 (use kernel regression). <br>
In the 1st edition, these are: the problem written below, and p195 #6, p195 #7, 
p195 #8, p195 #12 (use kernel regression).<br>

p243 #2 from the 2nd edition is the following:<br>
Use anova() to compare the two models:<br>
roller.lm = lm(depression~weight, data=roller)<br>
roller.lm2 = lm(depression~weight+I(weight^2), data=roller)<br>
Is there any justification for including the squared term?<br><br>
Homework 4 (due Wednesday Feb 14):
2nd edition: p284 #1, #3.<br>
On #1:
(a) Print output (coefficients, SEs, etc.)
(b) Interpret coefficients in 1-2 sentences.
(c) Plot fitted curve: logit(p) vs conc.<br>
1st edition: p223 #1, #3(a), #3(c).  See instructions above for #1.<br><br>