We now move on to one of the easily interpretable and most popular classifiers there are out there: the decision tree. Decision trees—which look like an upside down tree with the trunk on top and the leaves on the bottom—play an important role in situations where classification decisions have to be transparent and easily understood and explained. It also handles both continuous and categorical predictors, outliers, and irrelevant predictors rather gracefully. Finally, the general ideas behind the algorithms that create decision trees are quite intuitive, though the details can sometimes get hairy.
Figure 9.7 depicts a simple decision tree designed to classify motor vehicles into either motorcycles, golf carts, or sedans.