数据挖掘技术在职业卫生工程分析中的应用研究

1122(11,510300;21),,(OLAP),,;;;R13;D23A167124199(2007)0220070205ApplicationofDataMiningTechnologyonOccupationalHealthEngineeringMAIHai-ming,LUOHai-ming,CHANGHui-you,etal.GuangdongHealthInspectionInstitute,GuangzhouGuangdong510300,ChinaAbstractObjectiveToapplyhistoricaldataandexpertsexperiencetooccupationalhealthengineeringanalysisbyusingDataMiningtechnologysoastocomplementtheinsufficiencyoftraditionalmethodsandimprovethequalityofengineeringanalys2is.MethodsUsinghistoricaldatatosetupthedatabase,andcombiningOLAPandhazardassessmentmodalbasingonneuralnetworkintegrationtoexcavatethegeneralizationengineeringcharacteristicsandharmfulfactorscharacteristicsandhazardassess2mentregulations.ResultsThismethodcouldmakeafulluseoftheconnotativeregulationsofhistoricaldataandexpertsexperi2encetoimprovetheengineeringanalysisaccuracy.ConclusionTheapplicationofDataMiningtechnologyoffersanewwayforengineeringanalysismethodamelioration.KeywordsOccupationaldisease;Healthengineering;Hazardassessment;DataMining:(1962-),,,,:(OccupationalHealthEngi2neeringAnalysis),1(DataMining),,2,,,,1111:(7506):112,,Unknown,-1:,,,-1,,1,OheasDataWarehouse,RiskAnalysisEn2terpriseSituationHazard6EnterpriseSituation072007462ChineseJournalofPublicHealthEngineeringApr12007,Vol16No12©1994-2008ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.(OLAP)2,,,,,,,,,2115(RiskAssessmentofOccupationalHazard)(),,,,11511,MESLECMLS3,:R=max{Ri|i=1,2,,n}(1)Ri=max{Rijj=1,2,,m};i=1,2,,n(2)Rij=f(Kij,Pij,Lij,Mij,Eij,Sij};i=1,2,,n;j=1,2,,m(3),R;Rii;Rijij,1;n;m;Kijij;Pijij;Lijij;Mijij;Eijij;SijijLECMESMLS,,(M)(E)(S)(L)(k)(p)(R)(1)OheasDataWarehouseSLR:SOLAPR3L,OheasDataWare2house,:Oheas2DataWarehouseX={X1,X2,,Xn}(Xjm),2:,,;XjZj,XiZi,ZiXj;,,,()(CPT),ZiCPTPZi|Parents(Zi),,Parents(Zi)ZiZi,,Zn(Zi,,Zn):P(Z2,,Zn)=Kni=1PZi|Parents(Zi),PZi|parents(Zi)ZiCPT172007462ChineseJournalofPublicHealthEngineeringApr12007,Vol16No12©1994-2008ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.(M)12,,3,()(E)123,456(S)12,131456(L)1,2,3,,4,,,5(k)12(p)1100250993104945955(R)123(:=,==,,=,=),=11512,,ANTSEN523:2ANTSEN3ANTSEN,,,,,ANTSENAnt-ClassfierNeAntNeAntANTSEN,Ant-ClassfierNeAnt,615,OheasDataWarehouse,XX,,272007462ChineseJournalofPublicHealthEngineeringApr12007,Vol16No12©1994-2008ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.:IF(=)AND(M=2)AND(p=2)AND(E=2)AND(k=1)THEN(L=1)AND(S=1)AND(R=3)ELSEIF(=)AND(M=2)AND(p=2)AND(E=2)AND(k=2)THEN(L=4)AND(S=3)AND(R=1)ELSEIF(=)AND(M=2)AND(p=2)AND(E=2)AND(k=2)THEN(L=4)AND(S=3)AND(R=1)MESLkp:IF(L=1)AND(M=2)AND(p=2)AND(E=2)AND(k=1)AND(S=1)THEN(R=3)ELSEIF(L=4)AND(M=2)AND(p=2)AND(E=2)AND(k=2)AND(S=3)THEN(R=1)RART(riskassessmentruleta2ble)(2)2XXRARTk=1M=1M=2M=3k=2M=1M=2M=3LSEpRLSEpRLSEpRLSEpRLSEpRLSEpR112234322143221(1),R=max(R,R,R)=3()2,1,,,2,,,,,{=,=,},,,,MESLECMLS,3,4,,311M1:,2003:5026212JiaweiHan1MichelineKamber1DataMining:ConceptsandtechniquesM1Singapore:Elsevierlnc,2006:1032186131M1:,2004:1442212141M1:,2006:119212415HomikKM,StinchcombeM,WhiteH1MultilayerfeedforwardnetworksareuniversalapproximatorsJ1NeuralNetworks,1989,2:359236616DorigoM,ManiezzoV,ColorniA1Positivefeedbackasasearchstrate2gyR1TechnkalReport91-016,DipartimentodiElettronica,Po2litecnicodiMilano,IT,199117DorigoM1LearningandNaturalAlgorithma(inItalian)R1Ph1D1t2hesis1DipartimentodiElettronica,PolitecnicodiMilano,IT,199218ColorniA,DorigoM,ManiezzoV1Distributedoptimizationbyantcolo2nies:InproceedingsofthefirsteuropeanconferenceonartificiallifeM.Amsterdam:ElsevierPublishing,1992:13421421(76)372007462ChineseJournalofPublicHealthEngineeringApr12007,Vol16No12©1994-2008ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.(WAI)J1,2000,17(2):12812,,,1J1,2000,17(2):7027213,,,1J1,2000,9(2):171217414,,,1J1,2006,32(5):275227815,,,1J1,2004,33(2):130213316,,,1J1,2000,17(4):193219517,,,1J1,1996,15(5):296229812007202225()(73)9ChuanChen,YouqingChen1Antcolonyalgorithmbasedselectiveneu2ralnetworkensembleJ1InternationalConferenceonComplexSystemsandApplication,2006,6:16218110DorigoM,ManiezzoV,ColorniA1TheAntSystern:OptimizationbyacoIonyofcooperatingagentsJ1IEEETransactionsonsysterns,man,andcvbernetcs-partB,1996,1:29241111ChuanChen,YouqingChen1NeuralNetworkEnsembleBasedAntColonyClassificationRuleMiningR1FirstInternationalconferenceonlnnovativeComputing,InformationandControl(ICICIC),Beijing,China,VIII,2006:30112DietterichTG1Machinelearningresearch:FourcurrentdirecdonsJ1AIMagazine,1997,18:972136113BreimanL1BaggingpredictorsJ1MachineLearning,1996,2:1232140114HansenLK,SalarnonP1NeuralnetworkensemblesJ1IEEEtranspatternanalysisandmachineintelligence,1990,10:99321001115SollichP,KroghA1Leamingwithensembles:howoverfittingcanbeusefuI1In:TouretzkyD,MozerM,HasselmoMeds1AdvancesinNeu2ralInformationProcessingSystems8,CambridgeM1MA:MITPress,1996:190219612007203220(),(),,,,:59P:,:,:21,:410007,:0731-5602072(),13627318523,E-mail:y