期刊网_中国期刊网

Robust designs for models with possible bias and correlated errors

简介：Thispaperstudiesthemodel-robustdesignproblemforgeneralmodelswithanunknownbiasorcontaminationandthecorrelatederrors.ThetrueresponsefunctionisassumedtobefromareproducingkernelHilbertspaceandtheerrorsarefittedbytheqthordermovingaverageprocessMA(q),especiallytheMA(1)errorsandtheMA(2)errors.Inbothsituations,designcriteriaarederivedintermsoftheaverageexpectedquadraticlossfortheleastsquaresestimationbyusingaminimaxmethod.Acaseisstudiedandtheorthogonalityofthecriteriaisprovedforthisspecialresponse.Therobustnessofthedesigncriteriaisdiscussedthroughseveralnumericalexamples.
标签：误差模型稳健设计再生核希尔伯特空间移动平均过程偏差最小二乘估计

全文阅读

全文阅读

Hierarchical Mixture Models for Zero-inflated Correlated Count Data

简介：Countdatawithexcesszerosareoftenencounteredinmanymedical,biomedicalandpublichealthapplications.Inthispaper,anextensionofzero-inflatedPoissonmixedregressionmodelsispresentedfordealingwithmultileveldataset,referredashierarchicalmixturezero-inflatedPoissonmixedregressionmodels.AstochasticEMalgorithmisdevelopedforobtainingtheMLestimatesofinterestedparametersandamodelcomparisonisalsoconsideredforcomparingmodelswithdifferentlatentclassesthroughBICcriterion.AnapplicationtotheanalysisofcountdatafromaShanghaiAdolescenceFitnessSurveyandasimulationstudyillustratetheusefulnessandeffectivenessofourmethodologies.
标签：

全文阅读

全文阅读