I am trying to use proc_df to process my training data, and using max_n_cat to have the data one hot encoded.
However, how can i do the same one hot encoding for the test set?
I can’t set the same max_n_cat when calling proc_df on test set, as the test set may not having all the category as in training set (and vice versa).
So what’s the best way to deal with it? Combine the test and training and call proc_df and then split them apart?