Knowledge discovery from a data set on dementia through decision forest

Decision tree is one of the most popular classifiers used in a wide range of real world problems for both prediction and classification (logic) rules discovery. A decision forest is an ensemble of decision trees, and it is often built for predicting class values more accurately than a single decision tree. Besides improving predictive performance, a decision forest can be seen as a pool of logic rules with great potential for knowledge discovery. However, an standard-sized decision forest usually generates a large number of logic rules that a user may not able to manage for effective knowledge analysis. In this paper, we propose a novel, problem (data set) independent framework for extracting those rules that are comparatively more accurate as well as reliable than others. We apply the proposed framework on rule sets generated from two different decision forest algorithms from a publicly available data set on dementia and compare the subsets of rules with the rules generated from a single J48 decision tree in order to show the effectiveness of the proposed framework.