[R-sig-ME] Help with glmmADMB error messages

Ben Bolker bbolker at gmail.com
Wed Jan 7 16:44:27 CET 2015


The problem here is that there are some combinations (probably in the
interaction terms, which aren't showing up in your error message due
to truncation) that are missing from your data.

https://stat.ethz.ch/pipermail/r-sig-mixed-models/2014q1/021510.html
https://stat.ethz.ch/pipermail/r-sig-mixed-models/2012q4/019499.html

discuss this issue.

http://rpubs.com/bbolker/6311 shows a worked example of how to find
the problematic columns, and a bit of information on how to work
around the problem.

  Side question: I presume you're using glmmADMB rather than lme4
because you need to fit a negative binomial (or zero-inflated) model?


On Wed, Jan 7, 2015 at 10:33 AM, Aoibheann Gaughran <gaughra at tcd.ie> wrote:
> Hello, I have a global model that seems to be working (albeit with a subset
> of my dataset, not my full dataset). There are 9 explanatory variables, 2
> random variables and three interaction terms. However I can only get it to
> run with two of the interaction terms. The third (habitat:month) causes
> this error message which Ive web searched but cant seem to find an
> explanation for.
>
> Error in glmmadmb(field_count ~ age + sex + habitat + month + soil + slope
> + : rank of X = 139 < ncol(X) = 142
>
> Habitat has 10 levels and month has 12. Perhaps this is too much for the
> model to cope with? Can you shed any light on this problem?
>
> Also, does you know why my model breaks down the more data I include? At
> the moment Im including 10,000 rows of data, but I need to analyse ~45,000
> rows (and counting).
>
> Many thanks,
>
> --
> Aoibheann Gaughran
>
> Behavioural and Evolutionary Ecology Research Group
> Zoology Building
> School of Natural Sciences
> Trinity College Dublin
> Dublin 2
> Ireland
> Phone: +353 (86) 3812615
>
>         [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models



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