Imputation missing data

Imputation missing data

ID:40081385

大小:558.91 KB

页数:16页

时间:2019-07-20

Imputation missing data_第1页
Imputation missing data_第2页
Imputation missing data_第3页
Imputation missing data_第4页
Imputation missing data_第5页
资源描述:

《Imputation missing data》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、CHAPTER25Missing-dataimputationMissingdataariseinalmostallseriousstatisticalanalyses.Inthischapterwediscussavarietyofmethodstohandlemissingdata,includingsomerelativelysimpleapproachesthatcanoftenyieldreasonableresults.WeuseasarunningexampletheSocialIndicatorsSurvey,a

2、telephonesurveyofNewYorkCityfamiliesconductedeverytwoyearsbytheColumbiaUniversitySchoolofSocialWork.Nonresponseinthissurveyisadistractiontoourmaingoalofstudyingtrendsinattitudesandeconomicconditions,andwewouldliketosimplycleanthedatasetsoitcouldbeanalyzedasiftherewer

3、enomissingness.AftersomebackgroundinSections25.1–25.3,wediscussinSections25.4–25.5ourgeneralapproachofrandomimputation.Section25.6discussessituationswherethemissing-dataprocessmustbemodeled(thiscanbedoneinBugs)inordertoperformimputationscorrectly.MissingdatainRandBug

4、sInR,missingvaluesareindicatedbyNA’s.Forexample,toseesomeofthedatafromfiverespondentsinthedatafilefortheSocialIndicatorsSurvey(arbitrarilypickingrows91–95),wetypecbind(sex,race,educ_r,r_age,earnings,police)[91:95,]Rcodeandgetsexraceeduc_rr_ageearningspoliceRoutput[91,]

5、13331NA0[92,]21237135.001[93,]23240NA1[94,]113423.001[95,]131240.00NAInclassicalregression(aswellasmostothermodels),Rautomaticallyexcludesallcasesinwhichanyoftheinputsaremissing;thiscanlimittheamountofinformationavailableintheanalysis,especiallyifthemodelincludesmany

6、inputswithpotentialmissingness.Thisapproachiscalledacomplete-caseanalysis,andwediscusssomeofitsweaknessesbelow.InBugs,missingoutcomesinaregressioncanbehandledeasilybysimplyin-cludingthedatavector,NA’sandall.Bugsexplicitlymodelstheoutcomevariable,andsoitistrivialtouse

7、thismodelto,ineffect,imputemissingvaluesateachiteration.Thingsbecomemoredifficultwhenpredictorshavemissingvalues.Forexample,ifwewantedtomodelattitudestowardthepolice,givenearningsanddemographicpredictors,thenthemodelwouldnotautomaticallyaccountforthemissingvaluesofearni

8、ngs.Wewouldhavetoremovethemissingvalues,imputethem,ormodelthem.InBugs,regressionpredictorsaretypicallyunmodeledandsoBugsdoesnotknow

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。