Abstract: The application of the rough set theory in the discretization of the decision table is studied, and a noval discretization algorithm based on the rough set theory is presented. The algorithm of this paper is based on the importance of condition attributes. Firstly, the iner-information between condition attributes and decision attributes is used to measure the importance of condition attributes, according to which the condition attributes are sorted in a descending order. Secondly, all break points of every condition attributes are examined and the redundant ones are climinated. Finally, each value in the decision table is replaced by a number representing the break point, and then the decision table is discretized.

Key words: rough set, decision table discretization, data mining