I'm currently working with pregnant women data. Given that the same woman could have multiple pregnancies over the years, I tend to use GEE to obtain odd ratios of my variables of interest.
Now say I am interested in a classification problem with ~ 120 variables (whether the woman will have a preterm or not). Of course I cannot do this with GEE. I wanted to explore some models such as XGBoost, SVM or RF; but I am wondering if they can take repeated IDs into consideration. I was tempted to include LR on the list, but it does not seem right to do so given how data is correlated. That is why I am restricting my list to the three mentioned methods.
I did some research but it doesn't seem clear to me. Could someone help me out ?