数据挖掘和数据融合中的几个问题

《数据挖掘和数据融合中的几个问题》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。

1、SeVeraIprobIemsindataminingandinfbrmationfusionYueShihOng(InstituleofIndustrjalProcessControI，ZhejiangUnjVersity)AbstraCtDataminIngismostpromisIngand仉OurishingfrOn村erjndatabasesystem．Datamining．alsopopuIarIyreferredloasK-仃ow』bdged『S∞ve，y『仃Da招base(KD0)．istheautOmatedconVeniente)(trac

2、tiOnofpanems怕presenlingknOwIedgeimpIIc川ysto怕din

3、argedatabase，datahouses，andOthermassiveInformation怕pOsitoriesOurcapabilitiesofbothgene怕髓ngandcO¨ect

4、ngdatahavebeenincreasing阳pidlyintheIastseveraIdecades．C0nt—bu“ngfactorsIncIudethewidesp陀aduse0fbarcodesformDstcommerclalproducts．thecom

5、puterizatIOnOfmanybusiness．scjent嘀c。and90vemmenttransactions，andadvancesindataco

6、IectiontooIsrangIngfromscannedte)ctandimagepIa怕nTlstOsateⅢteremotesensingsystems．InaddⅢ0n．pOpuIaruseOftheWOndWidewebasa910ba

7、informatiOnsystemhas们OOdedusw

8、that怕mendousamOuntOfdataandInformation．ThisexpI

9、OsivegrO、ⅣthInstoreddatahaSgeneratedanurgentneedfornewtechniquesandautOmaledtOOIsthatcanInteI¨gentIyassjslusIntransforminglhevastamOuntsofdataintousefUIInforma“0nandknOwledae．CIuste—ngismaIntoOIIndalam．njng．ClusleringaIgOmhmstrytoparfitiOnasetOfunlabeIedinpulvectorsintoanumberofsubs

10、ets(cIuSters)suchthatdataInthesamesubsetaremo陀simjlartoeachOtherthantodatainothersubsets．TherearetwOkindsofunsupervIsedcIusteringa190rithms：fuzzy、悖rsuscnsp．CnspcIusteringassignanyvectort0uniquecIusterwh¨efuzzycIuSteringaIlowsthevectort0beIOngt0severaIcluste晤；ndjfFe怆ntmembersh岫degree

11、s，ThestudydeveIopsseveralfastcnspclushjnn9apprOachessuchascOVe卜based，gnd—based．andkemel—basedOnestOcha¨engetheseopenprobIems：hOwcanacIustenngapproachassessthosearbit怕ry-shapedanddensily-skewedcIuste噶?AIsO，hOwOne∞nobtaInafastapproachtocIusteringinIinearcompIexity?Inaddition，fuzzycIus

12、tenngandappro×imationpossessspecIaIposiaonindataminingwhentheun

13、abeledvectorshaveuncertainUVVeprobe峙clustenngmechanismandapp盱cabIe鼎eIds．ChOquetfuzzyintegralhasbecomemO怕andmOrepopularfusjonapprOach．Asbelngbasedonthefuzzymeasure．hOweve‘twomainproblemsansewhenusing仙zzymeasures．npractIc

14、alappIi明tIon．Oneist