search

wbonat / mcglm

Fitting multivariate covariance generalized linear models
GNU General Public License v3.0
21 stars 5 forks source link

Runtime based on number of response variables? #15

Closed akmorrow13 closed 5 years ago

akmorrow13 commented 5 years ago

Is there a maximum number of response variables that this is meant for? I tried running with 100, and it did not complete. What is the maximum this has been used for?

wbonat commented 5 years ago

Hi, Actually, it is a really difficult question. It will depend on the type of model you are fitting, on your computer, C and Fortran libraries that you have installed for your R. For the standard R installation and personal computer, probably 100 response is too much. For 100 response only the covariance structure has 100 + 100*99/2 parameters to be estimated.
You can use the option verbose = TRUE to show the progress of the fitting to have some idea how it is going on.

akmorrow13 commented 5 years ago

Thank you for your response @wbonat ! I will try running with verbose.