数据流上概念漂移的检测和分类.pdfVIP

Abstract Abstract IIlthe of information and age，thedevelopmentcommunication，computer network USthechanceto andstore amountsof technologygive capture large data, whichresultinthedatatakean tofind and explosivegrowth．Howpatterns，trends anomaliesinthese builda modelfromthese oneofthe datas，and simple datas，is intheinformation isan greatchallenges age．Dataminingtechnology important toaddressthese thereis new approach challenges．Inrecentlyyears a kindof of withthe information includecredit applicationappeared development technology cardfraud SenSordataand net． detection，networksecuritymonitoring powersupply Such modelischaracterizeddataarrives andin application by continuous，orderly real definesuchdataformasdatastreamwhichiscontinuousand time，scholars infiniteordered ofdata． potentially sequence Classificationis an taskinthedata domainandthe important mining characteristicsofthedatastreamsmodelhas new for brought traditional challenges classification tousethefast volumeofthe technology．How arrival，overwhelming dataandthe driftasthe tobuildamodel concept