Aroughset-basedmethodforupdatingdecisionrulesonattributevalues’coarsening
andre ning
HongmeiChen,TianruiLi*,IEEESeniorMember,ChuanLuo,Shi-JinnHorng,IEEEMember,
andGuoyinWang,IEEESeniorMember
Abstract—RuleinductionmethodbasedonRoughSetTheory(RST)hasreceivedmuchattentionrecentlysinceitmaygenerateaminimalsetofrulesfromthedecisionsystemforreal-lifeapplicationsbyusingofattributereductionandapproximations.Thedecisionsystemmayvarywithtime,e.g.,thevariationofobjects,attributesandattributevalues.ThereductionandapproximationsofthedecisionsystemmayalteronAttributeValues’CoarseningandRe ning(AVCR),akindofvariationofattributevalues,whichresultsinthealterationofdecisionrulessimultaneously.Thispaperaimsfordynamicmaintenanceofdecisionrulesw.r.t.AVCR.Thede nitionofminimaldiscernibilityattributesetisproposed rstly,whichaimstoimprovetheef ciencyofattributereductioninRST.Then,principlesofupdatingdecisionrulesincaseofAVCRarediscussed.Furthermore,theroughset-basedmethodsforupdatingdecisionrulesintheinconsistentdecisionsystemareproposed.ThecomplexityanalysisandextensiveexperimentsonUCIdatasetshaveveri edtheeffectivenessandef ciencyoftheproposedmethods.IndexTerms—Inconsistentdecisionsystem,Roughsettheory,Incrementallearning,Decisionrule,Attributereduction,Approximations
1INTRODUCTION
OUGH
SetTheory(RST),animportantmethodforclassi cationproblems[35],hasbeenwidelyusedinintelligentinformationprocessing[29],machinelearningandimageprocessing[20][34].If-thendeci-sionrulesdeducedbyRSTarerelatedtotwokindsofworks,i.e.,attributereductionandcomputationofapproximations.AttributereductionisanimportantapplicationofRST[2][4][12][23][42],whichaimstoremovetheunnecessaryattributesfromthedecisionsystemwhilekeepingitsclassi cationability(orotherproperties).Certainanduncertaindecisionrulesareinduceddirectlyfromapproximationsofthedecisionclasses.Attributereductiondecidestheantecedentofdecisionrules.Approximationsareusedtodeducedecisionrulesfromdifferentequivalenceclassesindifferentregions.Inreal-lifeapplications,thedecisionsystemS=(U,A,V,F,G)evolvesdynamically,i.e.,
R
and a multi-view decision method based on decision-theoretic rough set ...In information table, R can be induced by a subset of attribute A ? C...
Rough-Set-Based Attribute Dependencies: Foundations and Bio-Medical Data Applications_专业资料。维普资讯 http://www.cqvip.com 第2 5卷第2期 200 6年 6...
2014Attribute Reduction for Set-Valued Ordered Fuzzy Decision System_林学_...Firstly, we propose a dominance relation-based rough fuzzy set model in ...
Variable precision rough set for multiple decision attribute analysis_专业资料。A variable precision rough set (VPRS) model is used to solve the multi-...
Lin proposed a method for GrC based on ...decision classes on the universe U [40, 41]. ...Based on attribute values of objects, one may ...
2014 Color Image Segmentation Based on Decision-Theoretic Rough Set Model_信息与通信_工程科技_专业资料。今日推荐 157份文档 ...
decision from complex information system, a new rough set model based on ...(2) If a is a quantitative attribute and Va just contains discrete values...
Improved Rough Set Algorithms for Optimal Attribute Reduct_专业资料 暂无评价
银行信贷管理--Research of Credit Risk Measurement Based On Compatibility Rough Set and Fuzzy Set_金融/投资_经管营销_专业资料。Research of Credit Risk ...
Feature subset selection based on mahalanobis distance: a statistical rough set method_专业资料。维普资讯 http://www.cqvip.com ACAD JJXTU 0801:41 20,2...
热门文档