[R-sig-ME] Maximizer failed

[Ammara Mahmood]

I tried extra.args="-ndi 60000" and still got the message to increase the calls. I increased the calls to 80000 and got the same error, however, at 90,000 and 100,000 the program freezes and R runs forever and does not respond.

Could you please suggest another solution.

Thanks


I am trying to use the Glmmadmb to model truncated negative
binomial with random effects in the count and zero part. The models
run fine when i I do not include the random effects. The data is on
page views per individual at a news paper website over a period of
1 month, hence I want to include person specific random effects.
And I create the group Id by person.

So I have been able to run the following successfully:


output <- glmmadmb(tot_pv~   time_last_visit + friends_residual+
number_friends + like_counter + avg_gap_activity + dum_local +
dum_sport + dum_more + dum_tvzap  + number_like +
tot_daily_visits_o + tot_daily_visits_reg + female + age  ,
data=(subset(facebook_merge,tot_pv>0)), zeroInflation=FALSE,
family="truncnbinom1")

However when I include a random effect on each individual in the
data set I get the following error message:

output <- glmmadmb(tot_pv~   time_last_visit + friends_residual+
number_friends + like_counter + avg_gap_activity + dum_local +
dum_sport + dum_more + dum_tvzap  + number_like +
tot_daily_visits_o + tot_daily_visits_reg + female + age ,
(1|group_id) , data=(subset(facebook_merge,tot_pv>0)),
zeroInflation=FALSE, family="truncnbinom1")

Error message

The function maximizer failed (couldn't find STD file) In addition:
Warning message: running command './glmmadmb -maxfn 500 -maxph 5
-noinit -shess' had status 1

Need to increase the maximum number of separable calls allowed to
at least 20001 Current value is 20000 Use the -ndi N command line
option


I am currently working on a MAC.

Could you please help me solve this issue?


 Well, try adding the argument

extra.args="-ndi 60000"

to your call and see if it works (if you get another complaint,
increase the
value further).

 If you have any more questions along these lines, could you please send
them via r-sig-mixed-models at r-project.org<mailto:r-sig-mixed-models at r-project.org>?

 Ben Bolker