CODEWITHSUNDEEP
r
swihart
swihart

Reputation: 2738

glmmadmb error: maximum number of dependent variables of 200 exceeded

I needed to fit a model that had about 500 fixed effects. 

nb.xx.ff <-glmmadmb(count~Pint-1, 
                    data=dnoNM,
                    family="nbinom")

And was given the following error:

## maximum number of depdendent variables of 200 exceeded 
## use gradient_structure::set_NUM_DEPENDENT_VARIABLES(int i);
## to increase the number of dependent variables
## Error in glmmadmb(count ~ Pint - 1, data = dnoNM, family = "nbinom", zeroInflation = TRUE,  : 
##   The function maximizer failed (couldn't find STD file)
## changed working directory to /spin1/users/swihartbj/projects/_SMFA/Pxs_2015_ctrls_only/Rnw/Final/_ilmn5 
## removed temp directory /scratch/Rtmp/Rtmp4iDtvH/glmmADMB2384f4030d4a1

Is there a way to increase this from the R call?

Upvotes: 1

Views: 182

Answers (1)

swihart
swihart

Reputation: 2738

Yes.

Add "-ndv 3000" to extra.args, as recommended here. As long as the number of fixed effects is the only issue and the number (3000 in this example) listed is bigger than the number of fixed effects, this should work:

nb.xx.ff <-glmmadmb(count~Pint-1, 
                    data=dnoNM,
                    family="nbinom",
                    extra.args="-ndv 3000")

Upvotes: 1

Related Questions