A Novel Extension of Rough Set Model in Incomplete Information System

The purpose of this paper is to investigate the incomplete information system in which all unknown values are considered as "do not care" conditions by rough set technique. Based on the non-symmetric similarity relation for classification analysis in such information system, we propose the concept of valued similarity relation to show the degree that an object is similar to another one. The fuzzy rough set technique is employed to form rough approximations in terms of valued similarity relation. Further on the problem of inducing decision rules from incomplete decision system is addressed. Some numerical examples are employed to substantiate the conceptual arguments.