Spatial econometrics introduction英文学习教学教案.ppt

《Spatial econometrics introduction英文学习教学教案.ppt》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、IntroductiontoAppliedSpatialEconometricsAttilaVargaDIMETICPécs,July3,2009PrerequisitesBasicstatistics(statisticaltesting)Basiceconometrics(OrdinaryLeastSquaresandMaximumLikelihoodestimations,autocorrelation)EUPatentapplications2002OutlineIntroductionThenatureofspatialdataModellingspaceExplo

2、ratoryspatialdataanalysisSpatialEconometrics:theSpatialLagandSpatialErrormodelsSpecificationdiagnosticsNewdevelopmentsinSpatialEconometricsSoftwareoptionsSpatialEconometrics„Acollectionoftechniquesthatdealwiththepeculiaritiescausedbyspaceinthestatisticalanalysisofregionalsciencemodels”LucAn

3、selin(1988)IncreasingattentiontowardsSpatialEconometricsinEconomicsGrowinginterestinagglomerationeconomies/spillovers–(GeographicalEconomics)DiffusionofGIStechnologyandincreasedavailabilityofgeo-codeddataThenatureofspatialdataDatarepresentation:timeseries(„timeline”)vs.spatialdata(map)Spati

4、aleffects:spatialheterogeneityspatialdependenceSpatialheterogeneityStructuralinstabilityintheformsof:Non-constanterrorvariances(spatialheteroscedasticity)Non-constantcoefficients(variablecoefficients,spatialregimes)Spatialdependence(spatialautocorrelation/spatialassociation)Inspatialdataset

5、s„dependenceispresentinalldirectionsandbecomesweakerasdatalocationsbecomemoreandmoredispersed”(Cressie,1993)Tobler’s‘FirstLawofGeography’:„Everythingisrelatedtoeverythingelse,butnearthingsaremorerelatedthandistantthings.”(Tobler,1979)Spatialdependence(spatialautocorrelation/spatialassociati

6、on)Positivespatialautocorrelation:highorlowvaluesofavariableclusterinspaceNegativespatialautocorrelation:locationsaresurroundedbyneighborswithverydissimilarvaluesofthesamevariableEUPatentapplications2002Spatialdependence(spatialautocorrelation/spatialassociation)Dependenceintimeanddependenc

7、einspace:Time:one-directionalbetweentwoobservationsSpace:two-directionalamongseveralobservationsSpatialdependence(spatialautocorrelation/spatialassociation)Twomainreasons:Measurementerror(dataaggregation)SpatialinteractionbetweenspatialunitsModellingspac