Rule extraction from medical data without discretization of numerical attributes

Abstract

Association rule mining is a popular technique used to find associations between attributes in a dataset. When using deterministic algorithms, if the attributes have numerical values the usual approach is to discretize them defining proper intervals. But the discretization can notably affect the quality of the rules generated. This work presents a method based on a deterministic exploration of the interval search space without a previous discretization of the numerical attributes. It has been applied to medical data from an atherosclerosis study. The quality of the obtained rules seems to support this method as a valid alternative for this kind of rule extraction.