Quick Maximum Distribution Reduction in Inconsistent Decision Tables

Attribute reduction is a key issue in rough set theory, and this paper focuses on the maximum distribution reduction for complete inconsistent decision tables. It is quite inconvenient to judge the maximum distribution reduct directly according to its definition and the existing heuristic based judgment methods are inefficient due to the lack of acceleration mechanisms that mainstream heuristic judgment methods have. In this paper, we firstly point out the defect of judgment method proposed by Li et al. [15]. After analyzing the root cause of the defect, we proposed two novel heuristic attribute reduction algorithms for maximum distribution reduction. The experiments show that proposed algorithms are more efficient.