This exercise will be an excellent introduction to tree-based methods. I recommend applying this method to any supervised learning method because, at a minimum, you'll get a better understanding of the data and establish a good baseline of predictive performance. It may also be the only thing you need to do to solve a problem for your business partners. An example I can share was where the marketing team tasked me to try and reverse-engineer a customer segmentation done by an external vendor nearly two years in the past. We had the original survey data and the customer segment labels, but no understanding of how the data drove the segmentation.
Well, I just used the methods described in this section, and we could predict ...
