In this chapter, we introduced a new machine learning algorithm, decision trees, which we can use for marketing analytics in order to better understand the data and draw insights on customer behaviors. We discussed how decision tree models are different from logistic regression models, which you learned about in the previous chapter. You saw that decision tree models learn the data by partitioning the data points based on certain criteria. We also discussed the two measures that are frequently used when growing decision trees: the Gini impurity and entropy information gain. Using either of these measures, decision trees can grow until all of the nodes are pure, or until the stopping criteria are met.
During our programming exercises in Python and R, we used the bank marketing dataset from the UCI Machine Learning Repository. We started our programming exercised by analyzing...
