Filled The Default Data Based On Rough Set Theory And The Variable Domain Of Rule-based Reasoning

Rough set is proposed by Z.Pawlak in 1982, it is new mathematical tool which deals with fuzzy and uncertain knowledge, it is widely used in finding knowledge, date mining etc, so study gradually mature.Rough set is based on classification, it regards knowledge as classification of data, the main idea of rough set uses known knowledge in knowledge base to describe uncertain and inexact knowledge, the most differentiation between rough set and other theories to deal with uncertain and inexact knowledge is it need not provide any information for needing processed date excluding , so it is more objective to deal wit uncertain problem.Rough set is defined by upper and lower approximation, so it can define accuracy of approximation. Rough set can solve important problem of classification, it induce decision of problem and classification rules by knowledge reduction.The paper studied basic properties of rough set by upper and lower approximations. Some approaches which can fill missing data are introduced to incomplete information system. Based on research of rough set theory, an algorithm to fill missing date values is proposed using the attributes difference of objects and the notion of rough similarity, the paper pointed out that the algorithm can fill much more missing data. An approach of reduction decision table is proposed to induce decision rules. From the viewpoint of variable universal an approximate rule mining inference approach based on rough set theory is proposed . The rule base is regulated by minor premise and analysis rules ,so the rules are reasoned .