Belief discernibility matrix and function for incremental or large data

Abstract

This paper proposes an incremental attribute selection method based on rough sets from partially uncertain and incremental or large decision system. The uncertainty exists only in the decision attributes (classes) and is represented by the belief function theory. The simplification of large or incremental uncertain decision table is based on computing possible reducts by the means of belief discernibility matrix and function under the belief function framework from two or more sub-decision tables. © 2013 Springer-Verlag.