Study On Attribute Reduction Based Rough Set Theory

In this article,at first,we introduce Rough set theory, attribute reduction based on Rough set and the classification of information system,reduction algorithms of attribute. In this paper, we discusses some properties of two relations on set-valued incomplete information system, and obtains a new methods of knowledge reduction .In this paper,upper approximation and lower approximation are defined from the ideas of set pair analysis. Because Mulitivalude information systems reflect the uncertainty of information,and SPA is one kind of analysis uncertainty systems. upper approximation and lower approximation are defined from the ideas of set pair analysis, there is the certain usability. Andαdistribution reduction ,αassignment reduction are presented subsequently that to be allowed according to the people sameness degree which obtains to the identical request its needs, enabled the multiple-valued information system the application scope to obtain has further promoted. Then, we discover that reduction algorithm carries on two kinds of each characteristic the union, may obtain a better result . In this article, a combination reduction algorithm that based on characteristics of information entropy and discernibility matrix is studied .we proofed that the algorithm can attain balance between accurate degree and complex degree.