ABSTRACT
The focus of this chapter is on supervised rule-based machine learning techniques. Section 7.1 shows how decision trees can be mapped to a set of rules. An example demonstrates that it may be possible to simplify the original set of mapped rules. Section 7.2 outlines a fundamental covering rule algorithm. The JRip(RWeka package) covering rule algorithm is then applied to a customer churn dataset. Section 7.3 demonstrates an efficient technique for generating association rules. The Apriori(RWeka package) association rule function is then utilized to find interesting relationships in a customer database of grocery store purchases. In Section 7.4, the spotlight is on Rattle, a graphical user interface supporting many of the preprocessing, modeling, and evaluation methods discussed throughout the text. Rattle’s interface is used to generate production rules with rpart, model customer churn with the randomForest function, and generate association rules with the apriori(arules) function. Several end-of-chapter exercises provide practice-building rule-based models.
