[R] conditional regression with mgcv

[Simon Wood]

My guess is that the model has identifiability problems and that this is 
then causing a problem (not caught properly) in the model fitting 
optimizer. Is there any chance you could send data that produces the 
problem (off list) and I can try it out (I will only use any data for 
this investigation, of course - but if the data is confidential then any 
suitably messed up version that also fails would be just as useful).

best,
Simon

On 31/03/17 16:58, Dean Force wrote:
> Hello,
>
>
> As a part of a larger project, I am trying to run a conditional logistic
> regression to look at whether maternal age is implicated in the risk of
> developing gestational diabetes. I am using a matched case-control design,
> where mothers with GDM were individually matched with up to 6 controls
> based on several parameters.
>
>
> I run the following model:
>
>
> model <- gam(gdm ~ s(maternal_age, bs="cr") + strata(risk_set) +
> as.factor(district) + as.factor(riskfactor1)+as.factor(riskfactor2), data =
> dt, family=cox.ph(), weights = wt)
>
>
>
> weights are defined as 0 for censoring, 1 for event, and each subject has
> one event/censoring time and one row of covariate values. In total there
> are 1000 cases, matched to 5500 controls, so there are 1000 risk_set that I
> define as strata.
>
> When running the model, I keep getting the following error: “Error in
> xat[[i]] : subscript out of bounds”. Am I doing something wrong?
>
> Using mgcv_1.8.
>
>
>
> Thank you!
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.


-- 
Simon Wood, School of Mathematics, University of Bristol BS8 1TW UK
+44 (0)117 33 18273     http://www.maths.bris.ac.uk/~sw15190