In this chapter, we tackled the KDD Cup 2009 challenge on customer-relationship prediction, where we implemented the data pre-processing steps, addressing the missing values and redundant attributes. We followed the winning KDD Cup solution, studying how to leverage ensemble methods using a basket of learning algorithms, which can significantly boost the classification performance.

In the next chapter, we will tackle another problem addressing the customer behavior, that is, the analysis of purchasing behavior, where you will learn how to use algorithms that detect frequently occurring patterns.