Abstract. This paper presents a new philosophical view and methodology for data mining. A framework of explanation oriented data mining is proposed and studied with respect to association mining. The notion of conditional associations is adopted, which explicitly expresses the conditions under which an association occurs. To illustrate the basic ideas, the theory of rough sets is used to construct explanations.
We propose a new framework of explanation-oriented data mining by adding an explanation construction and evaluation phase to the data mining process. While traditional approaches concentrate on mining algorithms, we focus on explaining mined results. The mining task can be viewed as unsupervised learning that searches for interesting patterns. The construction and evaluation of mined patterns can be formulated as supervised learning that builds explanations. The proposed framework is therefore a simple combination of unsupervised and supervised learning. The basic ideas are illustrated using association mining. The notion of conditional association is used to represent plausible explanations of an association. The condition in a conditional association explicitly expresses the plausible explanations of an association.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.