{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

homework5

homework5 - 3(10 points Agresti Problem 3.18 a-c Note An...

This preview shows page 1. Sign up to view the full content.

Stat 665 (Spring 2011) Kaizar Homework 5 Due at the beginning of class, Friday May 6 . Aim of Homework : To practice computing with GLMs. Reading : An Introduction to Categorical Data Analysis, 2nd edition: Chapter 3. Exercises : 20 points total 1. (5 points) Agresti Problem 3.11, b and c. You may find this code useful: y=c(8,7,6,6,3,4,7,2,3,4,9,9,8,14,8,13,11,5,7,6) x=rep(c(0,1), c(10,10)) wafers=data.frame(y,x) 2. (5 points) Use the data in Agresti Problem 3.12 to construct a model where the link and linear components are: log ( μ i ) = α 1 + β 1 z i You may find this code useful: wafers\$z=rep(c(0,1,0,1),c(5,5,5,5)) (a) Report the prediction equation and interpret ˆ β 1 . (b) Compare this model to the one you found in (1) above. Which model do you prefer?
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 3. (10 points) Agresti Problem 3.18, a-c. Note: An o±set term is a known (not estimated) part oF the linear portion oF the model, as describedin Section 3.3.5. The “o±set” argument to the glm() Function in R might be useFul here. ²or example, you might want to use code similar to: glm(Y~X + offset(log(attendance)) The output oF this code will have estimates For an intercept and a coe³cient For X, but not For the o±set term. (IF you attended the Statistics Department seminar on Thursday, April 28, the speaker used a similar “o±set” to account For the population in the counties in Wisconsin when creating models For the number oF cancer cases observed in each county.) 1...
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online