[R-sig-ME] lme4 upgrade broke my old code?
bbolker at gmail.com
Wed Mar 18 21:26:13 CET 2015
-----BEGIN PGP SIGNED MESSAGE-----
On 15-03-17 02:24 PM, Stuart Luppescu wrote:
> On Tue, 2015-03-17 at 13:27 -0400, Ben Bolker wrote:
>> I think that starting with version 1, lme4 was not compatible
>>> the lower version code. I am afraid that you will have to
>>> calculate the old models again with the newer lme4.
>> There is some help here. If you install the lme4.0 package (see
>> the README at https://github.com/lme4/lme4 ) you can at least
>> use convert_old_lme4() to update lme4 pre 1.0 to lme4.0 ... We
>> don't currently have a function to convert stuff that old to
>> modern lme4 (lots of the internal structures have changed), but
>> if you want to badly enough you can use lme4.0 ...
> That's what I was afraid of. Thanks Marko and Ben. I'm rerunning it
> now -- takes a long time with 2.5 million observations.
Even in the absence of my more complicated suggestion (trying to
populate the merMod object with already-computed stuff), I would make
sure you are
(1) starting from the previously computed best-fit values
(2) using the known tricks for optimizing speed -- in particular I
would use control=lmerControl(optimizer="nloptwrap"), as the default
bobyqa optimizer uses tolerance criteria that are probably stricter
than they need to be.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
-----END PGP SIGNATURE-----
More information about the R-sig-mixed-models