Research Of Knowledge Discovery Methods Based On Rough Set Theory

In this thesis, we study deeply knowledge discovery methods based on rough set theory .Some available methods are summarized and some new methods are presented. All in all, the main important research is as follows:Firstly, we introduced the discernibility matrix and a reduct algorithm based on this matrix, latter, we studyed these information systemes which has some noises. In this situation,reducts may become meaningless or not appropriate for classification. So,we proposed two approximate reduct methods, one based on contingency matrix, the other based on variable precision discernibility matrix.Secondly, we present a new approach to selection of attributes for construction of decision tree based on rough set theory. Because of some existing rough set based methods are not applicable for large dataset, we present a classification method, which is equivalent to rough set based classification methods, but is scalable and applicable for large data sets. The proposed method is based on lazy learning idea and Apriori Algorithm.Thirdly, traditional rough sets approach pursuits the fully correct or certain classification rules without considering other factors such as uncertain class labeling, importance of examples, as well as the uncertainty of the final rules. A general rough sets model, GRS, is proposed and a classification rules induction approach based on GRS is suggested.Finally, we used entropy to extend the rough set based notion of a reduct. Latter, an approach of learning classification functions by genetic programming is proposed for classification.