An algorithm of attribute frequency reduction based on discernibility matrix

This paper expounds the basic conceptions of the rough sets theory, in which knowledge reduction is a very important topic. An algorithm which is based on the frequency of attribute is proposed in the discernibility matrix of an information system. The algorithm produces some result of all attribute reductions (when the attributes of two conditions appear at the same the frequency). Using the algorithm we can reduce the complexity of the original attribute reduction algorithm greatly and acquire the optimal decision-making rules of the information system.