Closed ThijsZwi closed 1 year ago
Problem solved:
I switched back to an earlier version of R that i was using (3.4) and the code and model ran perfectly fine. I updated yesterday to 3.6.3, so maybe that is the problem. Regardless, i will close this issue. Thanks for developing this package!
I updated to R version 4.0 and somehow this error keeps popping up. I have the newest version of pscl also installed. I updated since i saw that i was using PSCL 1.5.2. and a very simple zeroinfl() ran for ages without converging (even though it used to converge). Do you maybe have any idea what could be the issue?
At first i thought it was maybe due to the model itself, so i tried a different package for my zero inflation model (glmmTMB). This model did ran, so could it be that i have some issues with pscl itself?
Hello, I've been having the same issue! New to the pscl, but would be extremely helpful to know if someone managed to work around this!
This error appears when you include "EM = TRUE" argument. By removing it the function works. Also, could not find the description of the argument in the function documentation, so it might be from the earlier version of the package/function.
This issue has been solved in the most recent merge to the master branch, so I'll close the issue.
@maticjancic EM = TRUE is passed to thezeroinfl.control()
function which does document that argument. The zeroinfl function docs could be made clearer though by inheriting the function arguments, I agree.
Hello!
A few days ago i was running a different ZINB model which ran perfectly fine. However, today i wrote a different model, same variables, but somehow it doesn't run anymore and gives a strange error. Proposed model:
ex1zinb <- zeroinfl(X1 ~ Y1 + M1 | Y1 + M1, data = mydata, dist="negbin", EM= TRUE)
Error: Error in zeroinfl(Y1 ~ X1 + M1 | X1 + : object 'model_count' not found
So i tried the ZINB model that ran a few days ago, but suddenly i encounter the same problem, even though it are the same variables and dataset, so the model does not convert anymore. I had some issues with installing other packages, so i reinstalled the packages involved in the whole script. Then i tried to run it solely with the pscl library loaded but it gave me the same error. If i look into the pscl zeroinfl source code there is also an object called model_count, therefore i am thinking it is somehow related to this package. Curious thing is that if i delete EM=TRUE, it does somehow run, but only providing estimate outcomes and nothing else.
Do you have any clue what i possibly did wrong or if should somehow re-install pscl in a different way?
Thanks in advance! Thijs