Rule Extraction from Incomplete Decision System Based on Novel Dominance Relation

Some of the rules derived by DRSA (Dominance-based Rough Set Approach) from incomplete decision systems may contain unknown values. To overcome the shortcoming, this paper proposes a new DRSA, which is based on novel dominance relation. The adapted relation is distinguished from traditional dominance relation by two roles, a subject and a reference. The subject played by an object from the universe leads to one-way comparison with the reference done by a virtual object from the virtual space. By the approach, virtual objects without unknown values, instead of objects that may carry unknown values, are used to extract decision rules from incomplete decision system, which brings about the induced rules without any unknown values. A numerical example is employed to substantiate the conceptual arguments.