第03讲逐步回归.pptVIP专享

3逐步回归医用多元统计分析方法主要内容1衡量回归方程的标准2逐步回归3回归系数反常的原因4岭回归医用多元统计分析方法1衡量回归方程的标准复相关系数R校正复相关系数Radj剩余标准差总误差总回归SSSSSSSSR12总误差MSMSRpnnRadj1111122误差MSspxxxy21医用多元统计分析方法模拟数据X1X2X3X4YX1X2X3X4Y137261911.5166191410.21511403419.82410322619.8218291713.72211393825.31912153321.610717209.72711132722.3188342214.83210211519.12911282120.7178181611.71811163219.62610352319.41610153420.3146141810.6187231411.12813213425.52311292920.7199132918.72513414028.91210193819.3329121518.3238251715.63611371821.52811333224.7319251417.7219181915.32913143828.33514243429.81810113521.6医用多元统计分析方法例3.2资料的一切可能回归(24-1=15个)2adjRpxxxys21参数个数方程中变量R2CpAIC2X10.365290.3441319.787412834.0097.45623X20.915120.912292.64619354.7433.07465X30.051890.0202929.557574247.00110.29764X40.586000.5722012.906691839.0083.782623X1,X20.920780.915322.55491331.2232.86640X1,X30.375960.3329220.125702788.0098.91384X1,X40.993390.992930.213283.82-46.59486X2,X30.916010.910212.70887352.7434.73893X2,X40.922130.916762.51133325.1232.31589X3,X40.609070.5821112.607801737.0083.948024X1,X2,X30.921230.912792.63099331.1734.68250X1,X2,X40.993810.993140.206893.93-46.69119X1,X3,X40.993600.992920.213694.85-45.65645X2,X3,X40.923480.915282.55590321.0333.755905X1,X2,X3,X40.994010.993130.207425.00-45.77377医用多元统计分析方法2逐步回归1.前进法(step-up,forward-entryprocedure)2.后退法(step-down,backward-eliminationprocedure)3.逐步向前法(forwardstepwise)4.逐步向后法(backwardstepwise)医用多元统计分析方法前进法的基本思想选定一个标准。开始方程中没有自变量(常数项除外)按自变量对y的贡献大小由大到小依次挑选进入方程。(假设检验的P值越小贡献越大)每选入一个变量进入方程，则重新计算方程外各自变量对y的贡献。直到方程外变量均达不到入选标准，没有自变量可被引入方程为止。医用多元统计分析方法单因素回归分析结果(方程中只含有一个变量)方程变量回归系数标准误SEtP①x10.47929020.11534724.160.000②x22.5379590.141120117.980.000③x30.13814130.10779991.280.210④x40.48354820.07420476.520.000医用多元统计分析方法X2已经在方程中，增加哪个变量好？方程变量回归系数标准误SEtP⑤x22.40056100.168342914.260.000x10.07242910.05031871.440.161⑥x22.52119200.145965017.270.000x30.01845040.03336220.550.584⑦x22.29246200.204895211.190.000x40.07882890.04878401.620.117医用多元统计分析方法X2,X4已经在方程中，增加哪个变量好？方程变量回归系数标准误SEtP⑧x20.18068770.13123301.380.179x40.46806770.025761418.170.000x10.47420830.026344318.000.000⑨x22.26282000.210961310.730.000x40.08167620.04938131.650.109x30.02286100.03251600.700.488医用多元统计分析方法X2,X4,x1已经在方程中，是否增加X3？方程变量回归系数标准误SEtP(10)x20.17859820.13141741.360.185x40.46742740.025802718.120.000x10.47206870.026470917.830.000x30.00895980.00929560.960.344医用多元统计分析方法STATA的运行结果.swregyx1x2x3x4,pe(0.2)beginwithemptymodelp=0.0000<0.2000addingx2p=0.1169<0.2000addingx4p=0.0000<0.2000addingx1Source|SSdfMSNumberofobs=32---------+------------------------------F(3,28)=1497.53Model|929.4717223309.823907Prob>F=0.0000Residual|5.7929182628.206889938R-squared=0.9938---------+------------------------------AdjR-squared=0.9931Total|935.264643130.1698271RootMSE=.45485----------------------------------------------------------------y|Coef.Std.Err.tP>|t|[95%Conf.Interval]---------+-------...