A Covering Rough Fuzzy Set Model And Classification

Rough set theory and fuzzy set theory are two important methods to deal with the knowledge of uncertainty and incomplete in information systems.They are also important tools for data mining.Rough set theory,proposed by Poland scholar Pawlak in1982, use the upper and lower approximation set to approximate the sets by equivalence relation between objects in the domain.Fuzzy set theory,proposed by American scholar Zadeh in1965, uses membership function to describe the individual belongs to the fuzzy set membership degree,which emphasis on the individual fuzzy.Both have achieved great success on knowledge acquisition, pattern recognition, machine learning and other fields. In the range of applications, there are a lot of limitations on the classic rough set theory which is a division of information based on the equivalence relation.Ordinary fuzzy set theory can not contain both the information of individuals belonged to the fuzzy sets and the information of individuals opposed to fuzzy sets.So Zakowski extended the division of information to the coverage of domain.That is to extend the classical rough set theory to the Covering rough set theory.Atanassov presented the concept of intuitionistic fuzzy sets.Which made rough set theory and fuzzy set theory more extensive application.Rough set theory emphasizes the classification boundary,but fuzzy set theory emphasizes the fuzzy of individuals. Both have their focus and advantages.How to cover the approximate space and fuzzy set theory to fuzzy classification is an important research topic on combination of fuzzy sets and rough sets.Some researchers have researched on the approximation of the fuzzy set for covering approximation space,proposed some models.But they also have some flaws, needs to be improved.On this basis, a rough fuzzy set model based on the induction of coverage was proposed.This paper research the nature of the model and construct the fuzzy classifier.The main work is as follows:(1)Summarized the three existing covering rough fuzzy set models, which are in-depth analysed.(2)On the basis of the above models, a rough fuzzy set model based on the induced covering was proposed and the nature of the model was discussed.(3)As Fuzzy accuracy for the reference standard,The new model were compared with the three existing model by experiment and this model is extended to the covering rough intuitionistic fuzzy set model.(4)Based on the induced covering of rough fuzzy set model,the paper constructed the corresponding classifier and analyzed its advantages.