| In this chapter we first define the field of inductive machine learning and then describe Michalski’s basic AQ algorithm. Next, we describe two of our machine learning algorithms, the CLIP4: a hybrid of rule and decision tree algorithms, and the DataSqeezer: a rule algorithm. The development of the latter two algorithms was inspired to a large degree by Michalski’s seminal paper on inductive machine learning (1969). To many researchers, including the authors, Michalski is a “father” of inductive machine learning, as Lukasiewicz is of multivalued logic (extended much later to fuzzy logic) (Lukasiewicz, 1920), and Pawlak of rough sets (1991). Michalski was the first to work on inductive machine learning algorithms that generate rules, which will be explained via describing his AQ algorithm (1986). |