资源描述：
《Fuzzy sets in machine learning and data mining111》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。
1、FuzzySetsinMachineLearningandDataMiningEykeHullermeier•Philipps-Universit•atMarburgDepartmentofMathematicsandComputerScienceMarburg,Germanyeyke@informatik.uni-marburg.deAbstractMachinelearning,datamining,andseveralrelatedresearchareasareconcernedwithmethodsfortheautoma
2、tedinductionofmodelsandtheextractionofinter-estingpatternsfromempiricaldata.Automatedknowledgeacquisitionofthatkindhasbeenanessentialaspectofarti cialintelligencesincealongtimeandhasmorerecentlyalsoattractedconsiderableattentioninthefuzzysetscommunity.Thispa-perbrie
yr
3、eviewssometypicalapplicationsandhighlightspotentialcontributionsthatfuzzysettheorycanmaketomachinelearning,datamining,andrelated elds.Inthisconnection,someadvantagesoffuzzymethodsforrepresentingandminingvaguepatternsindataareespeciallyemphasized.1IntroductionAstheconce
4、ptionofintelligenceisinseparablyconnectedwiththeabilitytolearnfromexperienceandtoadapttonewsituations,itishardlyastonish-ingthatmachinelearninghasalwaysbeenconsideredasanintegralpartofthe eldofarti cialintelligence(AI).Infact,thekeyroleoflearningandadaptationisalmostse
5、lf-evidentfortheconnectionistapproachtoAI,sincerepresentingknowledgewith(arti cial)neuralnetworksbecomespossibleonlybytraining"suchnetworks.However,inviewofthefactthattheknowl-edgeacquisitionbottleneck"turnedouttobeoneofthekeyproblemsinthedesignofintelligentandknowle
6、dge-basedsystems,theimportanceofau-tomatedmethodsforknowledgeacquisitionhasalsobeenrealizedformoresymbol-orientedapproacheslikerule-basedsystems(forwhichexpertknowl-edgecouldinprinciplebeinjectedbyhand).Inanycase,machinelearningiscertainlyoneofthemostimportantsub eldso
7、fcontemporaryAI,regardlessoftheparticularfacet.1Aresearch eldcloselyrelatedtomachinelearningisthatofknowledgediscov-eryanddatamining.Asaresponsetotheprogressindigitaldataacquisitionandstoragetechnology,alongwiththelimitedhumancapabilitiesinanalyzingandexploitinglargeam
8、ountsofdata,this eldhasrecentlyemergedasanewresearchdiscipline,lyingattheintersectionofstatistics,machinelearning,dat
