Improvement of decision table attribute reduction algorithm based on Discernibility Matrix

The discernibility matrix, be put forward by Skowron in 1992, is one of effective knowledge acquisition approaches for knowledge representation system. This method is easy to understand and manipulate , but it may cause some error when we use it to calculate the relative core and reductions of inconsistent decisions, while redundant informations are also involved. To amend the limitation of the discernibility matrix way and to decrease computation complexity of its algorithm, the new inconsistent decision table attributes reduction algorithm based on discernibility matrix and fission tactic is proposed in this paper, which is suitable for whichever a KRS, and reducing algorithm computation complexity effectively. By comparing with the origin and other existing improving algorithm, the new attributes reduction algorithm based on discernibility matrix is showed more concision and more efficiency.