半参数部分线性变系数模型的统计推断

《半参数部分线性变系数模型的统计推断》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、？—－、．＇－合氏．ｌ心／ＩｌＩ＇？■？＇’．？＊、？、Ａ＇＞；，Ｉ－‘或‘、，＝＾、３＇：，山、＃諭—＊‘、，：．？了＾ＵＤＣ夺．＊？、－－、－．、？？－一、－＊＇戸．？，、‘＇—＇－－共、＇－＇－，＊＾‘：Ｖ．＼ｆ．．．＇尸－＾－Ｖ／，均書护―＇‘．．．，，—硕古緣隹讼＾；；，．；，矣；过；：＾－？、－．＇、‘＾＾－－＊／？〇＇、：二＊；＿ｒ一兴於：７？＊－－牛＾－’？．－‘

2、．．＇石．，、．■■、‘－’．＇在冥＊、．－－＾：－－异．．■‘一．——’、＾＇＾．：＇Ｖ＇户完：－－■－—？■＇梦一、■－二■？一安＇－＞．．■．．．．－＿－．左＞，＼＇、＿，。、‘、Ｎ论文题目：ＴｈｅＳ化ｔｉｓｔｉｃｓｉｎｆｅｒｅｎｃｅｏｆｓｅｍｍａｒａｍｅｔｉｉｃ－？和＼Ｖ＇Ｈ‘－、？、ｒｒｖ－＼Ｐａｒｔｉａ化Ｖｌｉｎｅａｖａｉｎｇｃｏｅｆｆｉｃｉｅｎｔｍｏｄｅｌ、坏、．、术、？、＇＇辛

3、节一文、专业名称：统计学？？＾人ｒ半参数估计／：．是研巧方向：－：张－，星：明学生姓名／一＞学号１１１１１１１００２．／：２０１３＾，‘Ｖ导师姓名：刘、夏自含天￡＿火：？３１．导师职称：教》、Ｂ１ＡＭｖ、令／，＇／’、－．－讀、．ｒ‘，心言杂‘＇ｈ’《ｖ０６年５月－：２１２８日：＿巧定滿时间，八、茂一－＇一—、Ｉ心；＾．＊’－－？－；．一？叫＊Ｖ＇？？、．？’、、？．’

4、＂‘－一．、？．，．．＇＇＾‘巧、二？－－祭阳Ｃ如中田巧皆＇＂？、．．．ｗ旅次乂■■ｒ；＇＊．－■、－■？：：巧＼＿！巧－：Ｖ＊－．、．－■？■－．二－．？；＼。．著？么六兴、京Thestatisticinferenceofsemi-parametricpartiallylinearvarying-coefficientmodelAbstractThispapermainlyusesthetechniqueofprincipalcomponentanalysi

5、s,krigingregressingimputation,GroupLassoandBootstraptosolvetheproblemsoftheestimationofsemi-parametricmodelsuchasmulti-collinearity,missingdataandgroupdata.Furthermore,weinvestigatetheestimationofsemi-parametricmodelonthesmallsampleconditionandprovidethecorre

6、spondingproprieties.Thedetailedworkisasfollows:Fortheproblemofmulti-collinearity,thispaperstudiestheback-fittingestimationofsemi-parametricpartiallylinearvarying-coefficientmodelsonthebasisofprincipalcomponentanalysis.Inordertoovercomethemulti-collinearityand

7、havebetterestimationefficiency,weapplyprincipalcomponentanalysistosemi-parametricpartiallylinearvarying-coefficientmodels.Thatistosay,weuseprincipalcomponentsasnewvariablesandthenweobtaintheestimatorsoforiginalparametriccomponentandnonparametriccomponentbythe

8、methodofback-fittingestimation.Furthermore,weprovidesomestatisticinferencesabouttheestimators.Fortheproblemofmissingdata,weinvestigatesaclassofestimationproblemsofthesemi-parametricmodelw