摘要
Datacollectedonthesurfaceoftheearthoftenhasspatialinteraction.Inthispaper,anon-isotropicmixingspatialdataprocessisintroduced,andundersuchaspatialstructureanonparametrickernelmethodissuggestedtoestimateaspatialconditionalregression.Undermildregularities,sufficientconditionsarederivedtoensuretheweakconsistencyaswellastheconvergenceratesforthekernelestimator.Ofinterestarethefollowing:(1)Alltheconditionsimposedonthemixingcoefficientandthebandwidtharesimple;(2)Differentlyfromthetimeseriessetting,thebandwidthisfoundtobedependentonthedimensionofthesiteinspaceaswell;(3)Forweakconsistency,themixingcoefficientisallowedtobeunsummableandthetendencyofsamplesizetoinfinitymaybeindifferentmannersalongdifferentdirectioninspace;(4)However,tohaveanoptimalconvergencerate,fasterdecreasingratesofmixingcoefficientandthetendencyofsamplesizetoinfinityalongeachdirectiona
出版日期
2002年04月14日(中国Betway体育网页登陆平台首次上网日期,不代表论文的发表时间)