基于聚類分析的航線旅客市場預測方法研究.pdf

1、南京航空航天大學碩士學位論文基于聚類分析的航線旅客市場預測方法研究姓名：王喆申請學位級別：碩士專業(yè)：交通運輸規(guī)劃與管理指導教師：夏洪山20090101基于聚類分析的航線旅客市場預測方法研究IIAbstractAirlinepassengermarketdemanalysisfecastingarethefoundationbasisofairlinesoperation.Accuratedemdatacanassistairlinest

2、omakescientificdevelopmentstrategyavoidblindnessofopeningnewairlineloweroperationcostsimproverevenue.Basedonanalysisofimpactfactsofairtransptationdemconventionalfecastingmethodsthispapermainlystudiedfecastingmethodfnewai

3、rlinepassengermarketdem.Firstlythispaperanalyzedimpactfactsofairpassengerdemthoughly.Basedonqualitativejudgmentwechosemainimpactfactsthroughgreycrelationanalysis.Theresultofgreycrelationanalysisshowedthattheinfluencedegr

4、eeofpercapitaGDPurbanresidentspercapitadisposableincomewerelagerthanGDPbecauseGDPisonlyageneralindicatChinahaslargepopulationairpassengerdemisdeterminedbyindividualchoicepeople’sspecificeconomicconditionsconsumptionpower

5、caninfluencetheirchoiceofairservicedirectly.Secondlyadvantagesdisadvantagesapplicabilityofconventionalairlinemarketfecastingmethodswereanalyzedevaluatedanentropybasedcombinedfecastingmethodwasintroducedtoenhancepredictin

6、gaccuracytheexampleverifiedthatthecombinedmethodcanpredictairpassengerdemmeaccuratelythananyconventionalmethod.Thirdlyafecastingmethodofnewairlinepassengermarketdemwasstudied.KMEANSclusteringmethodwasintroducedtoclassify

7、China’sdomesticairlinesbasedonfactanalysisafternewairlinetypewasdistinguishedloglinearmodelcouldbefoundedtopredictnewairlinepassengerdem.Thismethodcansolvetheproblemwiththenewairlineshtofhisticaldatathecoefficientsofvari

8、ablesofloglinearmodelareflexibilitycoefficientstheycanbeusedtostudyactersofairlinemarketdem.Airpts’servicerangescompetitivefactsweretakenintoconsiderationwhenclusteringindicatswerechosenasresultnumberofcompetitiveairptsa

9、veragedistancesofcompetitiveairptswereintroducedascompetitionindicatsothatairlinesampleschosencouldbemecomprehensive.Toavoiduncertaintyofclusteringanalysiscausedbydisparitybetweensamplesdecreasingvalueofeachindicatshould