Research On Attribute Reduction Method Of Incremental Information System Based On Dominance Relation

The database stores a large number of attributes unrelated to learning or recognition tasks.In order to eliminate the adverse effects of this redundant and irrelevant data information on our judgments,the attribute reduction of the information system has received extensive attention and research.Through reading related literature,we found that there have been few researches on attribute reduction of information systems based on dominance relations in the past,and there are many deficiencies.This paper mainly studies the attribute reduction of incremental information systems based on the dominance relationship.The specific research content of the article is as follows:First,we reviewed the knowledge of classical rough set theory and attribute reduction in information systems,at the same time,we do a certain research and discussion of the attribute reduction and increment problem of information systems under the classical rough set model,and then we introduce the related properties of the classical superior rough sets.Secondly,considering the deficiencies of the classic superiority relationship,this paper introduces several improved dominance relationships and we analyze the shortcomings of these improved dominance.At the same time,considering the importance of attributes and the difference between the same attributes will affect the dominant relationship between the objects,we have given a new dominance relationship,from this new “dominance relationship based on attribute value weighting”,we have established a fuzzy variable precision rough set model,and analyzes the relate properties of the newly established dominance relationship and the dominant variable precision rough set model,then we give an attribute reduction algorithm based on the newly established dominance relationship information system.In the last chapter of this article,we discuss the changes of the dominant class and the positive domain after the object changes in the information system.In this paper,we focus on the change of the dominant class of conditional attributes and decision attributes when a single object and multiple objects change,and give the calculation formula and the proof process.To a certain extent,it solves the problem of the dominant class and the positive domain after the change of the object set in the information system,at the same time,we give the incremental attribute reduction algorithm.