Introduction to Regression Analysis Chapter 4

Introduction to Regression Analysis Chapter 4

ID:40944771

大小:155.48 KB

页数:9页

时间:2019-08-11

Introduction to Regression Analysis Chapter 4_第1页
Introduction to Regression Analysis Chapter 4_第2页
Introduction to Regression Analysis Chapter 4_第3页
Introduction to Regression Analysis Chapter 4_第4页
Introduction to Regression Analysis Chapter 4_第5页
资源描述:

《Introduction to Regression Analysis Chapter 4》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、Chapter4ISimultaneousinferenceIRegressionthroughtheoriginIInverseprediction(leftasreading)IE ectsofmeasurementerrors(leftasreading)IChoiceofXlevels1/9Con denceintervals(CIs)andsimultaneousinferenceIFora(1)100%CIforoneparameterofinterest,e.g.95%CIfor1,

2、wecansay:IBeforeobservingthedata:Prob(Event1295%CI)=0.95IAfterobservingthedata:Weare95%con dentthat1isinits95%CIIFormultipleparameters,eachwithitsown(1)100%CIforoneparameterofinterest,e.g.a95%CIfor0anda95%CIfor1:IBeforeobservingthedata:Prob(fEvent0295

3、%CIgANDfEvent1295%CIg)=?IAfterobservingthedata:Weare?%con dentthat0isinits95%CI,AND1isinits95%CIIWhenconstructingcon denceintervalsformultipleparameters:IEachparameterhasitsowncon denceinterval,withitsownindividualcon dencelevel/coecient,e.g.1ICombin

4、edcon dencelevel=familycon dencecoecientIGoalinsimultaneousinference:Controlthefamilycon dencecoecient(bycontrollingtheindividualcon dencecoecients)2/9Bonferronijointcon denceintervalsIBonferroni:Whenconstructing2con denceintervals,wecanuse(11)and(

5、12)con dencelevelsforindividualintervals,with1+2=,toguaranteeanoverallcon dencecoecientof(1-)(proofonboard)IE.g.wecanuse95%con denceintervalsfor0and1,suchthatIbeforeobservingthedata:theprobabilityoftheeventthatatleastoneCIdoesnotcontaintheir"happeni

6、ngis0.10Iafterobservingthedata:weare90%con dentthateachCIcontainsitsIExtendBonferroni:Whenconstructingmcon denceintervals,wecanuse(1k)Pindividualcon dencelevels,withk=,toguaranteeanoverallcon dencelevelof(1-)IBonferronigivesconservativebounds:thefamil

7、ycon dencecoecientisATLEAST13/9SimultaneouspredictionforthemeanresponseIGoal:predictthemeanE(Yh)atmlevelsofX:fXh;h=1;:::;mgITwoapproaches:IBonferroniprocedure:use(1- =m)asindividualcon dencecoecientforeachXh:Y^hBsfY^hg;B=t(1;n2)2mIWorking-Hotell

8、ingprocedure(con dencebandfromCh2):pY^hWsfY^hg;W=2F(1 ;2;n2)IBothapproachesgiveconservativebounds,sochoosethetightestoneIInterpretationofa95%familycon dencecoecient?IA95%familycon dencecoecientformmeansmeansthatweare95%con denttha

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。