Algorithm For Attribute Reduction Of Concept Lattice And Rough Set-based Cluster Analysis To Explore

This paper is divided into two sections. The first section is mainly concerned with how to find all the absolute and relative reducts in an given information system. Reduction of attributes is one of the major concerns in the field of data mining. Firstly, a brief introduction is made to the theoretical basis, concept lattice. Secondly, by virtue of the analysis of the Hasse map of concept lattice, we present an new algorithm for finding all the absolute reducts, experimental results are made to validate the presented algorithm. And finally, a new algorithm for finding all the relative reducts is proposed.The application of Rough Set theory in designing clustering algorithm is discussed in the second section. It turns out with practical examples that the classification error can be greatly reduced by virtue of Rough Set theory methodology.