I do not get very well why the ID column needs to be ignored in the model; since it may be useful to goup by ID and know the total absent hours of each employe. So that, the probability of absent of each employee can be build.
Thanks for reaching out.
We’ve treated the absences of every individual separately, i.e. as separate events. We don’t group by ID, namely because this is something that can be ruled out during the regression (and that’s what happens). That’s why we keep the analysis focused on the events as opposed to the individuals.
Hope this helps.