Optimal decision rules acquisition for interval-valued ordered decision information systems

Abstract

The interval is viewed as basic knowledge granule and used to define the lower and upper approximations. This model is applied to interval-valued ordered decision information systems, and used to induce useful “at least and at most” decision rules. To obtain optimal decision rules, the concept of relative reducts of an interval is proposed, and the corresponding discernibility function is constructed for computing the relative ruduct.