摘要
Anewapproachtoknowledgeacquisitioninincompleteinformationsystemwithfuzzydecisionsisproposed.Insuchincompleteinformationsystem,theuniverseofdiscourseisclassifiedbythemaximaltoleranceclasses,andfuzzyapproximationsaredefinedbasedonthem.Threetypesofrelativereductsofmaximaltoleranceclassesarethenproposed,andthreetypesoffuzzydecisionrulesbasedontheproposedattributedescriptionaredefined.ThejudgmenttheoremsandapproximationdiscernibilityfunctionswithrespecttothemarepresentedtocomputetherelativereductbyusingBooleanreasoningtechniques,fromwhichwecanderiveoptimalfuzzydecisionrulesfromthesystems.Atlast,threetypesofrelativereductsofthesystemandtheircomputingmethodsaregiven.
出版日期
2010年06月16日(中国Betway体育网页登陆平台首次上网日期,不代表论文的发表时间)