北大暑期课程《回归分析》(Linear-Regression-Analysis)讲义PKU5.doc

北大暑期课程《回归分析》(Linear-Regression-Analysis)讲义PKU5.doc

ID:52524243

大小:263.00 KB

页数:9页

时间:2020-03-28

北大暑期课程《回归分析》(Linear-Regression-Analysis)讲义PKU5.doc_第1页
北大暑期课程《回归分析》(Linear-Regression-Analysis)讲义PKU5.doc_第2页
北大暑期课程《回归分析》(Linear-Regression-Analysis)讲义PKU5.doc_第3页
北大暑期课程《回归分析》(Linear-Regression-Analysis)讲义PKU5.doc_第4页
北大暑期课程《回归分析》(Linear-Regression-Analysis)讲义PKU5.doc_第5页
资源描述:

《北大暑期课程《回归分析》(Linear-Regression-Analysis)讲义PKU5.doc》由会员上传分享,免费在线阅读,更多相关内容在工程资料-天天文库

1、Class5,Page9Class5:ANOVA(AnalysisofVariance)andF-testsI.WhatisANOVAWhatisANOVA?ANOVAistheshortnamefortheAnalysisofVariance.TheessenceofANOVAistodecomposethetotalvarianceofthedependentvariableintotwoadditivecomponents,oneforthestructuralpart,andtheotherfo

2、rthestochasticpart,ofaregression.Todaywearegoingtoexaminetheeasiestcase.II.ANOVA:AnIntroductionLetthemodelbe.Assumingxisacolumnvector(oflengthp)ofindependentvariablevaluesfortheith'observation,.Thenisthepredictedvalue.sumofsquarestotal:because.Thisisalwa

3、ystruebyOLS.=SSE+SSRImportant:thetotalvarianceofthedependentvariableisdecomposedintotwoadditiveparts:SSE,whichisduetoerrors,andSSR,whichisduetoregression.Geometricinterpretation:[blackboard]DecompositionofVarianceIfwetreatXasarandomvariable,wecandecompos

4、etotalvariancetothebetween-groupportionandthewithin-groupportioninanypopulation:Class5,Page9Prove:(bytheassumptionthat,forallpossiblek.)TheANOVAtableistoestimatethethreequantitiesofequation(1)fromthesample.Asthesamplesizegetslargerandlarger,theANOVAtable

5、willapproachtheequationcloserandcloser.Inasample,decompositionofestimatedvarianceisnotstrictlytrue.Wethusneedtoseparatelydecomposesumsofsquaresanddegreesoffreedom.IsANOVAamisnomer?III.ANOVAinMatrixIwilltrytogiveasimpliedrepresentationofANOVAasfollows:(be

6、cause)(inyourtextbook,monsterlook)SSE=e'eClass5,Page9(because,asalways)(inyourtextbook,monsterlook)IV.ANOVATableSOURCESSDFMSFwithRegressionSSRDF(R)MSRMSR/MSEDF(R)ErrorSSEDF(E)MSEDF(E)TotalSSTDF(T)Letususearealexample.Assumethatwehavearegressionestimatedt

7、obey=-1.70+0.840xANOVATableSOURCESSDFMSFwithRegression6.4416.446.44/0.19=33.891,18Error3.40180.19Total9.8419Weknow,,,,.IfweknowthatDFforSST=19,whatisn?n=20=201.71.7+0.840.84509.12-21.70.84100-125.0Class5,Page9=6.44SSE=SST-SSR=9.84-6.44=3.40DF(Degreesoffr

8、eedom):demonstration.Note:discountingtheinterceptwhencalculatingSST.MS=SS/DFp=0.000[askstudents].Whatdoesthep-valuesay?V.F-TestsF-testsaremoregeneralthant-tests,t-testscanbeseenasaspecialcaseofF-tests.Ifyouhavedifficultywi

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。