An Clustering Algorithm Based on Rough Set

Abstract

Based on rough set theory, the paper proposed a clustering algorithm to deal with the quality and efficiency of clustering algorithm. By use of the consistency of condition attributes and decision attributes in the information table, the algorithm introduced a formula of attributes importance to reduce the redundant attributes. According to the data super-cube and entropy, the algorithm discretized the information table from global angle to local angle. Due to the set feature vector and set dissimilarity, the algorithm can cluster data just by scanning the information table only one time. The result of experiment indicates that the algorithm is efficient and effective