A value supplementation method for case bases with incomplete information

Abstract

In this paper we present a method for supplementing incomplete cases with information from other cases within a case base. The acquisition of complete and correct cases is a time-consuming task, but nevertheless crucial for the quality and acceptance of a case-based reasoning system. The method introduced in this paper uses association rules to identify relations between attributes and, based on the discovered relations we are able to supplement values in order to complete cases. We argue that using these related attributes when retrieving supplementation candidates will yield better results than simply picking the case with the highest global similarity. The evaluation of the method is carried out using four different publicly available case bases. © 2009 Springer Berlin Heidelberg.