R.C.H. CHENG ANALYSIS OF SIMULATION ANALYSIS OF SIMULATION OUTPUT BY RESAMPLING

R.C.H. CHENG ANALYSIS OF SIMULATION ANALYSIS OF SIMULATION OUTPUT BY RESAMPLING

ID:37659041

大小:199.78 KB

页数:8页

时间:2019-05-27

R.C.H. CHENG ANALYSIS OF SIMULATION ANALYSIS OF SIMULATION OUTPUT BY RESAMPLING_第1页
R.C.H. CHENG ANALYSIS OF SIMULATION ANALYSIS OF SIMULATION OUTPUT BY RESAMPLING_第2页
R.C.H. CHENG ANALYSIS OF SIMULATION ANALYSIS OF SIMULATION OUTPUT BY RESAMPLING_第3页
R.C.H. CHENG ANALYSIS OF SIMULATION ANALYSIS OF SIMULATION OUTPUT BY RESAMPLING_第4页
R.C.H. CHENG ANALYSIS OF SIMULATION ANALYSIS OF SIMULATION OUTPUT BY RESAMPLING_第5页
资源描述:

《R.C.H. CHENG ANALYSIS OF SIMULATION ANALYSIS OF SIMULATION OUTPUT BY RESAMPLING》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库

1、R.C.H.CHENG:ANALYSISOFSIMULATIONANALYSISOFSIMULATIONOUTPUTBYRESAMPLINGR.C.H.CHENGFacultyofMathematicalStudiesUniversityofSouthamptonSouthamptonSO171BJAbstract:Bootstrapresamplingisanextremelypracticalandeffectivewayofstudyingthedistributionalpropertiesofsimulationoutputwhenthisissubjectt

2、orandomvariation.Thepurposeofthispaperistoshowthatmanyproblemsofsensitivityanalysis(SA)andvalidationcanbeverysimplyhandledbythismethod.Weconsiderclassicalnon-parametricversionofthebootstrapandshowthatitcanbeappliedinanaturalwayinsimulationstudieswithoutbecomingimmersedincomplicatedstatis

3、ticalmethodology.ThemainmessageofthepaperisthatthebootstrapmethodshouldbethestatisticalmethodoffirstresortinSAandvalidationstudies.Keywords:Bootstrapping,sensitivityanalysis,validation,metamodels1INTRODUCTIONq=q(y)Bootstrappingisawell-understoodmethodforstudyingthedistributionofstatistic

4、alvariablesbycalculatedfromy.Atypicalexampleiswhereqisresampling[DavisonandHinkley1997,Chernickanaveragevalue1999].Thislatterreference,withits84page1bibliography,indicatesthehugeinterestintheq=Ξy(t)dt.Ttopic.Thusthecommoninitialreactionofmanynon-expertsthatitisasuspectmethodwhereoneTechn

5、icallyqisthusavectorfunctionalofy,buttriestoget'somethingfornothing',isquiteweneednotconcernourselveswiththis;allweunfounded.Itistruethatforfinitesamplesizesanneedtonoteisthatqisarandomvariablesothatapproximationisinvolved;howeverashasbeenitsstudyreducestoconsideringitsdistributionalampl

6、ydemonstratedboththeoreticallyandinproperties.practice,themethodisoftenespeciallyeffectivewhensamplesizesaresmall,whenotherwell-Inthenextsectionweconsiderinmoredetailtheusedandmuchreliedonmethodshavepropertiesssettingtypicallyencounteredinasimulationthataredubious[Hall,1992].model.Sectio

7、n3describesthebootstrapmethodingeneralterms,whilstinSection4itisappliedThepurposeofthisarticleistopointoutthetotwobasicproblemsinsensitivityanalysisandgeneralityofapplicationofbootstrappingbyinvalidation.Section5discussestwo,moreconsidering,inaunifiedway,howitcanbeadvance

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。