A Comparative Study Of Multi-scale Rough Sets And Concept Lattices Via Rule Acquisition

Currently,with the progress of the technology,the volume of data is gradually expanding.It has become a research hot spot to extract interesting information from a large amount of complex data.The idea of multi-scale has been introduced into rough set theory and concept lattice theory to realize multi-level knowledge discovery of information system and formal context.However,the existing researches on multi-scale concept lattices are not directly based on the binary formal context to discuss the multi-scale formal context.This brings some obstacles to some scholars who study the related problems of concept lattice from the viewpoint of the classical formal context.Meanwhile,rough sets and concept lattices are complementary to each other in knowledge discovery,granular description,uncertainty analysis and so on.Introducing the rule of multi-scale rough set into the theory of multi-scale concept lattice can expand the ability of multi-level knowledge discovery of concept lattice.It is also helpful for decision analysis of multi-scale data sets.In view of this,this paper gives the formal definition of multi-scale formal context based on the classical formal context,the difference and relationship between multi-scale rough set and multi-scale concept lattice rule extraction methods are also explored.The main results and innovations are as follows:(1)The methods of forward scaling and backward scaling are introduced,and the method of mutual transformation between information system and formal context are given.Based on the classical formal context,the definition of multi-scale formal context is given.This paper discusses the relationship between multi-scale information system and multi-scale formal context,and the relationship of implication rules from different scales.(2)The concrete steps and methods of transforming multi-scale decision information system into multi-scale decision formal context are proposed,and some related properties are discussed.The rule extraction methods between multi-scale rough set and scale concept lattice are compared and analyzed.