Research And Application Of Attribute Reduction Based On Rough Set Theory

Attribute reduction is one of important question in Rough set. It can eliminate unimportant and redundancies attribute of information systems.In the paper, the consistent sample set was defined, the equivalence relation between the consistent sample set and relative positive region is proven, and the attribute significance was defined based on boundary region. Moreover, a effective method of attribute reduction is presented, the main process is to calculate consistent sample set, extract condition relative attribute core, calculate relative attribute reduction set. In the attribute core calculation part, a theorem of estimate core attribute is given, in the attribute reduction part, a theorem of equivalence between reduction set and consistent sample set is given. The analysis indicated that the algorithm of calculate attribute core has low time complexities; the method of attribute reduction applies to consistent information system and inconsistent information system. The emluator result show that the validity of the algorithm were demonstrated.Lastly, in order to Rough set theory can applies to incomplete system, a new model base on k-equivalent tolerance relation is presented. The model has good flexibleness by performance analysis.