[R-sig-ME] convergence

Thierry Onkelinx thierry@onkelinx @ending from inbo@be
Fri Dec 21 16:29:49 CET 2018


Quasi complete separation is one of the things that might cause the
convergence problems. You'll need to investigate whether any combination of
your covariates causes (quasi) complete separation. If it is in the fixed
effect, you'll need to simplify the model. If it is a single random effect
level, then you can get away with it.

ir. Thierry Onkelinx
Statisticus / Statistician

Vlaamse Overheid / Government of Flanders
INSTITUUT VOOR NATUUR- EN BOSONDERZOEK / RESEARCH INSTITUTE FOR NATURE AND
FOREST
Team Biometrie & Kwaliteitszorg / Team Biometrics & Quality Assurance
thierry.onkelinx using inbo.be
Havenlaan 88 bus 73, 1000 Brussel
www.inbo.be

///////////////////////////////////////////////////////////////////////////////////////////
To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to say
what the experiment died of. ~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data. ~ Roger Brinner
The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of data.
~ John Tukey
///////////////////////////////////////////////////////////////////////////////////////////

<https://www.inbo.be>


Op vr 21 dec. 2018 om 16:22 schreef Souheyla GHEBGHOUB <
souheyla.ghebghoub using gmail.com>:

> Update on context:
> Aim: estimating the association between (1) responses accuracy and (2)
> time, groups, and 7 word-covariates.
>
> I included 2 covariates in the interactions with group & Time.
> Number of observations : 5800
>
> glmer(
>   *Response* ~ Characters + Band_freq + Imageability + Cognateness +
> Syllables + *Group*Time*PoS* + *Group*Time*Verbal_freq *
> + (Time|Subject) + (Time| Word),
> glmerControl(optimizer="bobyqa", optCtrl = list(maxfun = 2e5)), data =
> written.Form.data
>   family = 'binomial'
>
> On Fri, 21 Dec 2018 at 14:46, Souheyla GHEBGHOUB <
> souheyla.ghebghoub using gmail.com> wrote:
>
> > Should I "ignore" or get "too concerned" about this warning?
> >
> > Model failed to converge with max|grad| = 0.00542358 (tol = 0.001,
> > component 1)
> >
>
>         [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>

	[[alternative HTML version deleted]]



More information about the R-sig-mixed-models mailing list