The Information System Based On Dominance-disadvantage Relationship About Attribute Reduction

Firstly, the paper deals with the relationship between the property of the promoted Generalized approximate space (P(U),R) and the original Generalized approximate space (U,R). If (U,R) is quasi-ordering set, the condition of order homomorphism and iso-morphs between the two Generalized approximate spaces can be obtained. In dominance-disadvantage relationship information system, a new information system is produced by us-ing granulating method. Therefore, The dominance (disadvantage) relationship is converted to partial-ordering relationship. and the partial-ordering dominance-disadvantage relation-ship is promoted. The compatible set and reduction basing on dominance(disadvantage) relationship are in accordance with the original ones. Secondly, the paper introduces the concept of the distribution reduction in dominance-disadvantage relationship target information system.Using the distribution consistent set's equivalent condition and the discernibility matrix, the distribution reduction judgment theorem are proven based on dominance-disadvantage relation-ship. the conclusion is that the distribution reductions of the target information system basing on dominance-disadvantage relationship respec-tively are not identical. At last, with the rough set model basing on the dominance-disadvantage relationship,the definitions of precision reduction and upper approximate distribution reduction are acquired in the fuzzy target information system.The upper ap-proximate discernibility matrix and upper approximate reduction judgment theorem can also be acquired. According to the examples, the conclusion is that the precision reductions of the fuzzy target information system basing on dominance-disadvantage relationship re-spectively are different, so are the approximate reductions. This will be helpful to find out the knowledge of the information system basing on the dominance-disadvantage relation-ship.