[rkward-devel] local

[meik michalke]

hi,

am Donnerstag 05 August 2010 (09:21) schrieb Thomas Friedrichsmeier:
> well, I am somewhat curious as to the use-case for this.

funny, i've just used it myself to fix the inconsistent Wald test results 
(some got NaNs, others sane numbers) in the IRT plugin ;-) i've discussed this 
issue with the eRm devolpers, and they've addressed it in eRm 0.13. seems to 
be the default optimizer (nlm) that's causing it. if you run into this, you 
can now set the global object 

 fitctrl <- "optim"

to use optim() instead of nlm().

this worked for me, so for the sake of stable plugintests i've added it as an 
option to the Wald test dialog. if set to "optim", it now creates fitctrl in 
.GlobalEnv, runs Waldtest() and removes fitctrl again. maybe this will become 
a real option to the function (which i would prefer), but for the time being i 
think this should do for a solution.

you can check it out in the svn commits i did earlier today.


viele grüße :: m.eik

-- 
dipl. psych. meik michalke
institut f"ur experimentelle psychologie
abt. f"ur diagnostik und differentielle psychologie
heinrich-heine-universit"at 40225 d"usseldorf
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/rkward-devel/attachments/20100806/f1bdda4f/attachment.sig>