[R-sig-ME] Number of Iterations lme4

[Thierry Onkelinx]

Dear Jacob,

This might be a problem of (quasi)-complete separation. You have probably
ID's with only 0 values.
Another possibility is that the warning is a false positive. Have you tried
other optimizers? See ?glmerControl() for more information.

Best regards,

ir. Thierry Onkelinx
Instituut voor natuur- en bosonderzoek / Research Institute for Nature and
Forest
team Biometrie & Kwaliteitszorg / team Biometrics & Quality Assurance
Kliniekstraat 25
1070 Anderlecht
Belgium

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

2015-05-07 15:02 GMT+02:00 Jakob Gährken <jakob.gaehrken op uni-kassel.de>:

> Hello,
> I was asked to forward my question to this mailing list. I would be glad
> to get some hints regarding my questions below:
>
> I am an quit inexperienced R user but I try hard to learn fast. My
> problem is that while I try to run lme4 with the model:
> model1 = glmer(Survival ~ Kohorte+KJ+Saison+KAE+(1|ID), family =
> binomial(link = "logit"),
>
> data=phenos,control=glmerControl(optimizer="bobyqa",optCtrl=list(maxfun=1e6)))
> I always get the warning :
>
> 1:In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv,  :
>     Model failed to converge with max|grad| = 0.75662 (tol = 0.001,
> component 2)
> 2:In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv,  :
>     Model failed to converge: degenerate  Hessian with 1 negative
> eigenvalues
>
> I know that my data has the promblem that the binary response variable has
> a skewed distribution (1.933 "0" vs. 68.031 "1"). And that this could be
> the reason for the converging problem and that I might look for
> alternatives.(btw. do you consider some kind of minimum proportion for the
> "minor observation" in this case "0" (2.8%) when you fit a model e.g. for
> diseases with low incidence)
> But beside of the reason for the converging problem I am wondering because
> I tried to change the number of iterations using  :
> control=glmerControl(optCtrl=list(maxfun=...)).The used time which leads to
> the error message and also the  ...with max|grad| = 0.75662...are  always
> the same - no matter what value for maxfun I choose. So I am not sure how
> to control the number of iterations that are really done by R and whether
> the amount changes with different defined maxfun values in my case.
>
> Thank you for your time and looking forward hearing from you,
>
> Jakob Gährken
>
>
>
>         [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models op r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>

	[[alternative HTML version deleted]]