Knowledge reduction in interval-valued information systems

Abstract

In this paper, the concept of α -maximal consistent blocks is proposed to formulate the new rough approximations to an arbitrary object set in interval-valued information systems. The a -maximal consistent blocks can provide the simpler discernibility matrices and discernibility functions in reduction of interval-valued information systems. This means that they can provide a more efficient computation for knowledge acquisitions. Numerical examples are employed to substantiate the conceptual arguments.