2. We have a medical-related dataset. After using different models, we found that random forest works best. Now we want to build the final production classifier (using random forest). How would we divide the training data for the best model selection? (10) *