Decision Making by a Multiple-Rule Classifier: The Role of Rule Qualities

A rule-inducing learning algorithm yields a set of decision rules that depict knowledge discovered from a (usually large) dataset; therefore, this topic is often known as knowledge discovery from databases (KDD). Any classifier (or, expect system) then can utilize this decision set to derive a decision about given problems, observations, or diagnostics. The decision set (induced by a learning algorithm) may be either of the form of an ordered or unordered set of rules. The latter seems to be more understandable by humans and directly applicable in most expert systems, or generally, any decision-supporting one. However, classification utilizing the unordered-mode decision set may be accompanied by some conflict situations, particularly when several rules belonging to different classes match (are satisfied by, “fire” for) an input to-be-classified (unseen) object. One of the possible solutions to this conflict is to associate each decision rule induced by a learning algorithm with a numerical factor, which is commonly called the rule quality (An & Cercone, 2001; Bergadano et al., 1988; Bruha, 1997; Kononenko, 1992; Mingers, 1989; Tkadlec & Bruha, 2003). This article first briefly introduces the underlying principles for defining rules qualities, including statistical tools such as contingency tables and then surveys empirical and statistical formulas of the rule quality and compares their characteristics. Afterwards, it presents an application of a machine learning algorithm utilizing various formulas of the rule qualities in medical area.