异常数据挖掘算法分析与其在税务上的应用

《异常数据挖掘算法分析与其在税务上的应用》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、山东大学硕士学位论文ABSTRACTInthisarticleIrewillapplyclusteringandoutlierdetectionmethodondataoftax．Asthedatabasetechnologyhasbeenusedonrevenuewidely。revenuehasaccumulatedalargenumberofrowdata,whicharesavedinnatabasetolittleavail．Howtoabstractknowledgefromthesedataisthekeytas

2、kofDataminingtechnology．DataMiningisanewtechniquedevelopedfrom1980s．Itaimstoextracttheimplicit，unknown，andpotentiallyusefulknowledgefromvoluminous，non-complete，fuzzy，stochasticdata．Outlieranalysisisaimportantpartofdataminingresearch．Itspurposeistofindthe’smallpatter

3、ns’fromdataset．Anoutlierisanobjectthatisconsiderablydissimilarorinconsistentwiththeremainderofthedat山Thisisveryusefulinrevenue．Theoutlierinrevenuedatabasecouldbegeneratedbyaspecialmodeofproduction。alarge—scaletaxpayer，orevencriminality．^11oftheseareinspecialsupervis

4、ionofrevenue．Itisimportantforrevenuetofindthemquicklyandaccurately．TheoutlierdetectiontechnologyadaptedtorevenueisdiscussedinthisarticleFirstly。wedescribethebasicconceptsandmetho&Thenintroducethecommonlyobjectsandrepresentativeapplications．Westudyclusteringandoutlie

5、rdetectiontechnologyanddescribethecommonlyrules，andintroducesomeclusteringandoutlierdetectionalgorithms．Theresearchprocessandthecurrentsituationofoutlierdetectionarereviewed．Thealgorithmsofoutlierdetectionbaseddistance；density，deviationandhighdimensionareintroduce山T

6、hecontentofthesealgorithmsisanalyzed．Thedisadvantagesandadvantagesofthese山东大学硕士学位论文algorithmsarecompared．TheemphasisofthisarticleisusingODACDS(outlierdetectionalgorithmonContinuedDataSets)，oneofdensity-basedclusteringmethodtoanalyzethedataoftax．Thealgorithmcandiscov

7、erarbitraryshapeclustersandcalldistinguishnoise．Owingtothefeatureoftax，outlierdetectioncanbeusedwidelyinthefield．Forthedemandofrevenue，westudiedallkindsofalgorithmsaboutoutlierdetectionandClusterin$Onthebaseofstudyingtheclusteringalgorithmbasedoutlierfactor。webringf

8、orwardanoutlierdetectionalgorithmonContinuedDataSets．ThenewalgorithmsCanbeusetofindtheexcursionofthedata．Wefirstlyintroducetheconceptsofou