Using imprecise user knowledge to reduce redundancy in Association Rules

Abstract

Redundancy is a handicap in association rules. It becomes a limitation to use rules models in order to support the decision-making process. A technique based on user knowledge has been proposed recently , which aims at eliminating redundancy. However, it ignores the imprecise nature of knowledge. In this paper, the notion of knowledge redundancy is generalized and a method to propagate the user certainty over derivate rules is developed. Certainty factor models are used. Obtained results have shown a model reduction of 50% with previous knowledge below 3%. This method improves the eciency of