Rule evaluation model as behavioral modeling of domain experts

Abstract

In this paper, we present an experiment to describe behavior of a human decision on the rule evaluation procedure, which is a post-processing procedure in data mining process, based on objective rule indices. The post-processing of mined results is one of the key factors for successful data mining process. However, the relationship between transitions of human criteria and the objective rule evaluation indices has never been clarified as behavioral viewpoints. By using a method based on objective rule evaluation indices to support the rule evaluation procedure, we have evaluated the accuracies of five representative learning algorithms to construct rule evaluation models of the actual data mining results from a chronic hepatitis data set. Further, we discuss the relationship between the transitions of the subjective criteria of a medical expert and the rule evaluation models. © Springer Science + Business Media, LLC 2009.