Re: Model overfitting in ASREML-R manual?

From: Butler, David <David.Butler_at_DPI.QLD.GOV.AU>
Date: Tue, 8 Jan 2008 09:52:25 +1000

Kevin,

This is a typo in the manual - the model should also include 'season' in
the fixed effects, ie

m6 <- asreml(circ ~ x + season,
             random = .............)

With 'season' included the job runs to completion; presumably the
initial estimates for the us() model are inappropriate with 'season'
omitted.

I'll update the manual.

Dave.

-----Original Message-----
From: ASReml users discussion group [mailto:ASREML-L_at_AGRIC.NSW.GOV.AU]
On Behalf Of Kevin Wright
Sent: Saturday, 5 January 2008 6:14 AM
To: ASREML-L_at_AGRIC.NSW.GOV.AU
Subject: Model overfitting in ASREML-R manual?

In the ASREML-R manual, version 2, page 120, is the following model:

m6 <- asreml(circ ~ x,
             random = ~ us(link(~1+x), init=c(5,-0.01,0.0001)):Tree +
             spl(x) + spl(x):Tree,
             splinepoints = list(x =
c(118,484,664,1004,1231,1372,1582)),
             data = orange)

When I try to execute this, I get the following:

Equations: 42 (2 dense)
Initial update shrinkage factor: 0.316
     LogLik S2 DF

   -113.9692 56.5670 33 14:10:07 (3 component(s) constrained)
Warning: 2 1 NonPosDef US matrix modified ABORT Fault 2130 No residual
variation Finished on: Fri Jan 04 14:10:07 2008

Error: Abnormal termination
No residual variation

Can anyone confirm this error?
I hypothesize there has been a change in the dynamic linked code since
the manual was created.

My setup: Windows XP
asreml(): 2.00 Library: 2.00bi Run: Fri Jan 04 14:10:07 2008

Thanks,

K Wright

********************************DISCLAIMER****************************
The information contained in the above e-mail message or messages
(which includes any attachments) is confidential and may be legally
privileged. It is intended only for the use of the person or entity
to which it is addressed. If you are not the addressee any form of
disclosure, copying, modification, distribution or any action taken
or omitted in reliance on the information is unauthori
eflect the opinions
of the Queensland Government and its authorities. If you received
this communication in error, please notify the sender immediately and
delete it from your computer system network.
Received on Fri Jan 08 2008 - 09:52:25 EST

This webpage is part of the ASReml-l discussion list archives 2004-2010. More information on ASReml can be found at the VSN website. This discussion list is now deprecated - please use the VSN forum for discussion on ASReml. (These online archives were generated using the hypermail package.)