Attribute Reduction Algorithm Based on Discernibility Matrix of Skowron and Itemset Lattice

Attribute reduction algorithm for decision table based on Skowron´s discernibility matrix which is hardly decide the independence of the subset of condition attribute relative to decision attribute. In this paper, we firstly use item set lattice on Skowron´s discernibility matrix, and then search the item set lattice, at last use Boundary algorithm to get reduction sets. Experimental result indicates that the proposed method can get all reduction sets and don´t need to verify the problem of independence.