[6351] in s-news-athena
Monotonic regression with binomial errors
daemon@ATHENA.MIT.EDU (John Maindonald)
Wed Feb 1 14:50:19 1995
Date: Thu, 2 Feb 95 08:27:42 +1300
From: John Maindonald <john@maths.marc.cri.nz>
To: s-news@stat.wisc.edu
Does anyone have code for this?
I have been using the pool-adjacent-violators algorithm, followed
by the use of loess on the transformed proportions. This is a bit
of a kludge. Also I do not know how to get SEs of the predicted
values that I'd believe.
John Maindonald email : john@maths.marc.cri.nz
Hort+Research phone : 64-9-849-3660 Ext. 6707
P.B. 92169, Auckland fax : 64-9-815-4201
New Zealand
