移動(dòng)網(wǎng)格聚類分析及其在數(shù)據(jù)流管理中的應(yīng)用研究.pdf

1、合肥工業(yè)大學(xué)碩士學(xué)位論文移動(dòng)網(wǎng)格聚類分析及其在數(shù)據(jù)流管理中的應(yīng)用研究姓名：鄭盈盈申請學(xué)位級別：碩士專業(yè)：管理科學(xué)與工程指導(dǎo)教師：倪志偉20090301MobileGridClusteIingAnalysisandItsApplicationResearchinDatastreamManagementABSTRACTInrecentyears，agrowingnumberofapplicationenvironmentgenerateda

2、tastreamsSuchdataarecontinuous，orderly，rapidchanging，massiveHowtomineusefulinformationindatastreamhasattractedopenattentionClusteringiSanimportantdataminingmethodaswellastheresearchonapplyingclusteringtechnologiestodatas

3、treamsThetraditionalclusteringtechnologiescannotbeappliedtodatastreamsdirectly，SOitisworthresearchingdatastreamclusteringalgorithmHowtostoresuchdatashouldberesearchedwhenminingthedataThetraditionaldatabasecannotbeadapted

4、todatastreamsItiSnecessarytesearchnewdatastreammanagementsystemThecurrentresearchindatastreammanagementfocusedoncontinuousqueryhardlycombiningdatastreamminingwithdata—streammanagementThispaperstudiedthetraditionalmethods

5、ofdataminingandthecurrentdatastreamclusteralgorithms，thedatastreamclusteralgorithmMGDDSbasedonmobilegridanddensityispresentedThisalgorithmreducesthedefectsofcurrentdatastreamclusteralgorithmandimprovestheresultofprocessi

6、ngthenonsphericalcluster，whichcanhandleanytypeofclusterandenhancetheaccuracyoftheclusterItanalyzesthelimitationofthedatabasemanagementsysteminthispaper，thefeaturesofdatastreammanagementsystemunderthedatastreamenvironment

7、andthedistinctionbetweenDSMSandDBMSTheimprovedstructureofdatastreammanagementsystemispresentedwithreferencingonthecurrentdatastreamsystemcombiningwiththedatastreamminingAtlast，thepropertiesofthealgorithmwereanalyzedfocus

8、ingontheactualfieldofdatastreamandtheimplementationprocessofalgorithminnetworkintrusiondetectionwasgiven，examiningthecommonlydatasetKDDCup99withagoodresultaswellasauthenticatedthepropertiesofalgorithmKeywords：DataMining，