I will try to post .pdf copies of graphs used in lecture but not
included in the notes. This only includes computer generated
graphs. If (when?) I get really busy, graphs will be
posted late or not at all, so please do not rely on these.
Lecture on 5 Nov. rMSE and
rMSEP. Square root of MSE (fit to
current observations) and MSEP (prediction of new observations) for
1, 2, 3, ... 50 variables in the multiple regression. The true
model depends only on X1, X2 and X3. The remainder are noise.