Attribute Reduction on Distributed Incomplete Decision Information System

摘要

Attribute reduction is an important issue in rough set theory. This paper mainly studies attribute reduction of distributed incomplete decision information system (DIDIS). Firstly, the definition of rough set in DIDIS is developed. Next, an algorithm for attribute reduction of DIDIS is proposed. In the end, two groups of experiments are conducted to prove the effectiveness of the proposed method. The results show that our method can remove redundant attributes of DIDIS, and does not reduce the classification capability of the system. In addition, the results indicate that the change of data missing rate has weak effect on attribute reduction with the similarity relation, but strong effect on attribute reduction with the tolerance relation.