A Few Kinds Of Incomplete Information Systems’ Attribute Reduction

Attribute reduction of information system is an important part ofrough set theory, we can remove the unnecessary attributesduring the attributereduction. Then it can simplify the problem, keep basic information and save alot of work.Classical rough set theory is based on equivalence relation and incomplete in-formation system. In many practical problems, however, it is usually incomplete,such as incomplete information system, incomplete fuzzy information system,incomplete interval-valued information system, incomplete interval-valued fuzzyinformation system, incomplete intuitionistic fuzzy information system and so on.First of all, the definition of generalized incomplete intuitionistic fuzzy in-formation system is introduced based on incomplete intuitionistic fuzzy informa-tion system, and bring in the feature relation and structure discernibility matrixto discuss the attribute reduction of generalized incomplete intuitionistic fuzzyinformation system and generalized incomplete intuitionistic fuzzy decision infor-mation system.Then, we give the concept of diference degree based on tolerancerelation in incomplete information system, and use heuristic approximate re-duction algorithm to research the attribute reduction of incomplete informationsystem; At last, we structure dependence space based on the probability domi-nance relation of interval-valued information system and use dependence space model to reduct, and then define the extended probability dominance relation inthe incomplete interval-valued information system, we reduct the system whichexists both intersecting and containing attribute values in the incomplete interval-valued information system by using this relation to structure discernibility matrix.