The final visit information variable in this dataset is the length of the visit variable (LOV). However, the length of visit is determined only after the entire ED visit, and by that time, the decision whether to admit or discharge will have already been made. It is important to drop variables that won't be available during the time of the prediction, and for that reason, we must drop LOV. We do so as shown in the following code:

X_train.drop('LOV', axis=1, inplace=True)
X_test.drop('LOV', axis=1, inplace=True)

Now that we've finished tackling the visit information, let's move on to demographic variables.