aishaolang (3) [Avatar] Offline
#1
The callout "Log Odds" in section 1.2.2 says:
It turns out that the log(Odds) varies between -1 and 0.


It should be "... varies between -1 and 1."
aishaolang (3) [Avatar] Offline
#2
The arrows for New data and Answers in Figures 1.7, 1.12 and 1.13 point in different directions as compared to figure 2.1. I am inclined to think figure 2.1 has the correct flow; otherwise, the feedback loop for iterative improvements in figure 1.13 does not make sense.
dalila (2) [Avatar] Offline
#3
Just a note. A categorical variable with N levels should have N-1 binary variables.
438541 (1) [Avatar] Offline
#4
chapter 1 errata - German Credit Data unavailable
There appears to be a broken link. Footnote 5 on page 10 of Chapter 1 states that we can download the German Credit Data dataset from http://mng.bz/95r4 . This link successfully redirects to http://archive.ics.uci.edu/ml/datasets/Statlog+(German+Credit+Data) . Unfortunately, the top level domain ( http://archive.ics.uci.edu ) only provides an error message stating "Unable to connect" . Does anybody know of another way to access this data set? Thanks.