Classical-R, R-work, R-free
barani at deft.cc.purdue.edu
Wed Aug 28 01:21:55 EST 1996
Referring to R (WORKING set) as Rw, free R (10% TEST set)
as Rf and the classical R as Rc,
Rc (100% data) shows a marginally higher value than Rw.
If this is expected, then is it not proper to refer to Rw
and Rf more precisely than simply R and Rf where there is
a confusion between Rw and Rc? (particularly in publications)?
On the other hand, posting an article at midnight instead of
sleeping may be a bad idea :-)
ps: For those of you busy in the x-ray lab, there is a useful
and interesting article in Structure, Vol.4, 1996, p897
on checking your imagination over R-free.
More information about the X-plor