A new algorithm of attribute reduction based on fuzzy clustering

This paper proposes a new approach of attribute reduction for decision systems based on rough set and fuzzy clustering in order to avoid information loss resulted from the discretization of real valued condition attributes. In this paper, the fuzzy clustering technique is employed to obtain an optimal value which measures the inconsistency between condition attributes and decision attribute, and attribute reduction is performed to keep this optimal value. Finally, an example is employed to illustrate our idea.