Research For Rough Set Models Under Simi-larity Connnection Degree Tolerance Relation

This paper aims at those questions of dealing with date in set-valued infor-mation system,and comes up with the rough set model under similarity connection degree tolerance relation to solve some problems in reality. Firstly, this paper brings in the degree of connection and degree of similarity to set-valued information system under tolerance relation, then combines the threshold of degree of identity-α and threshold of degree of opposites-β, both of which belong to degree of connection, to adjust the influence degree for data and the roughness degree of decision-making and establish rough set model of tolerance relation of similarity connection degree. Secondly, it classifies the new model and compares the classi-fication results with those rough set models based on ordinary tolerance relation and limited tolerance relation proposed by other researchers. And analysis of the difference of these three classification results, which explains that the tolerance relations of new model proposed by this paper is effective for the classification of set-valued information system. Lastly, this paper uses this model to achieve the attributes reduction. Throughout the process of analyzing, this paper respectively discusses the complete and incomplete situation of set-valued information system combining with the application, and the results adequately explains the reliability and validity of this model put forward in this paper.