Then the models would not need to be combined in a list to be compared. I believe it would be OK to make this change to the function interface as the function is not present in the CRAN version of saemix.
I would also propose to stop if the requested likelihood is not available, as calculation the likelihood with the "gq" method can take a very long time (I am currently waiting for the results for my fits) and the user may not be prepared for that.
If one uses "ll" as suggested in the current documentation, the result for importance sampling is returned.
I would propose to change the interface to
Then the models would not need to be combined in a list to be compared. I believe it would be OK to make this change to the function interface as the function is not present in the CRAN version of saemix.
I would also propose to stop if the requested likelihood is not available, as calculation the likelihood with the "gq" method can take a very long time (I am currently waiting for the results for my fits) and the user may not be prepared for that.