The Research On Reduction Based On Rough Sets

Rough Set theory, which was proposed by Polish mathematician Pawlak.Z in 1982, is a mathematical tool to analyze uncertain and vague data and has been one of the challenging fields of computer science and technology. The Rough Set approach which applied to knowledge discovery, data reduction, decision support, classification and other fields has been proved very effective. The theory has been found in many interesting real-life applications.Data reduction is the key part of the Rough Set theory. Reduction of knowledge consists in removing superfluous attributes (values) in the knowledge base, in such a way that the elementary knowledge is preserved, so it can increase the knowledge quality and help people make decision. This dissertation focuses on the research of attribute reductions and value reductions, the contributions are as follows:(1) An introduction to Rough Set theory has been given. Existing algorithms of attribute (value) reduction have been described and analyzed.(2) A new attribute reduction algorithm ARIMC for both consistent and inconsistent decision tables is presented. This method also shows its advantages over the classical algorithm in less time consumption by using IDM and core properties as the optimization conditions.(3) An improved attribute value reduction algorithm AVRIMC for decision tables is presented. It utilizes the value core and absorptivity to optimize the matrix construction. Compared with other algorithms, this method shows its better performance by experiment.