[R-sig-ME] Data frame size limits in MCMCglmm?

[Stuart Luppescu]

Hello, I'm having problems running a simple ordinal outcome mixed
effects model, and I'm thinking it may be because of the size of the
dataset (or, it very may well be that I'm not specifying the model
correctly). I must confess to insecurity about how to specify the
priors. Here is the structure of the data frame (with columns not in
this model omitted). Note that there are more than 2.4 million rows. Is
that a problem?

 str(all.subj)
'data.frame':	2438922 obs. of  112 variables:
 $ gr10                  : num  0 0 0 0 0 0 1 0 1 0 ...
 $ gr11                  : num  0 0 0 1 1 1 0 0 0 0 ...
 $ gr12                  : num  1 1 1 0 0 0 0 1 0 0 ...
 $ tid                   : Factor w/ 14982 levels "........","A.D46607",..: 2 2 2 2 2 2 2 2 2 2 ...
 $ final.points          : Ord.factor w/ 5 levels "0"<"1"<"2"<"3"<..: 4 4 4 2 3 3 2 2 3 2 ...

Here are two attempts and their results:

glmm.uncond <- MCMCglmm(final.points ~ gr10 + gr11 + gr12,
                         prior=list(R=list(V=1, fix=1), G=list(G1=list(V=1, nu=0))),
                         random = ~tid ,
                         family = "ordinal",
                         nitt=100000,
                         data = all.subj)

Error: segfault from C stack overflow


 glmm.uncond <- MCMCglmm(final.points ~ gr10 + gr11 + gr12,
                         prior=list(R=list(V=1, nu=0), G=list(G1=list(V=1, nu=0))),
                         random = ~tid ,
                         family = "ordinal",
                         nitt=100000,
                         data = all.subj)


Process R segmentation fault (core dumped) at Fri Jan 18 12:53:49 2013

Here is my sessionInfo
R version 2.15.1 (2012-06-22)
Platform: x86_64-redhat-linux-gnu (64-bit)

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
 [7] LC_PAPER=C                 LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] foreign_0.8-52  MCMCglmm_2.17   corpcor_1.6.4   ape_3.0-6      
[5] coda_0.16-1     Matrix_1.0-10   lattice_0.20-13 tensorA_0.36   

loaded via a namespace (and not attached):
[1] compiler_2.15.1 gee_4.13-18     grid_2.15.1     nlme_3.1-107   
[5] tools_2.15.1   

and memory info.

 gc()
            used   (Mb) gc trigger    (Mb)   max used    (Mb)
Ncells   1196029   63.9    1835812    98.1    1710298    91.4
Vcells 678385919 5175.7 1777423119 13560.7 2114537169 16132.7

Any help will be appreciated.


-- 
Stuart Luppescu <slu at ccsr.uchicago.edu>
University of Chicago