An Efficient Method for Attribute Reduction in Incomplete Information Systems

Abstract

Attribute reduction is an important issue of data mining. In this paper a novel method based on rough sets is provided for attribute reduction in incomplete information systems. Through a transformation technique, an incomplete system is firstly converted into a new and simpler system and then reducts are obtained from the transformed system. It is proved by theorem that the transformed system has the same reducts as the previous one. Experiments show that the proposed method is more efficient on reduct computation of incomplete information systems.