A Fast Enumeration Algorithm for Attribute Reduction Based on Grouping

Abstract

When studying attribute reduction, we need to enumerate all the minimal reductions to prove the correctness of the attribute reduction algorithm. The explosion of attributes combination makes time complexity of the enumeration algorithm pretty high. An algorithm derived from grouping the rows of decision table is put forward to validate if an attribute set is reduction. Based on this algorithm, a fast algorithm for seeking core attribute is put forward too. And by optimizing the process of enumeration, the quantity of enumeration is reduced and the time of enumeration is reduced to acceptable scope.