《回归第四次作业统计一班07许林秀.doc》由会员分享，可在线阅读，更多相关《回归第四次作业统计一班07许林秀.doc（4页珍藏版）》请在金锄头文库上搜索。
1、（1） 建立y对x2x6的线性回归方程。由下面系数表可得y=5922.827+4.864x2+2.374x3-817.901x4+14.539x5-846.867x6Variables Entered/Removed(b)ModelVariables EnteredVariables RemovedMethod1X6, X3, X2, X4, X5(a).Entera All requested variables entered.b Dependent Variable: YModel SummaryModelRR SquareAdjusted R SquareStd. Error of t
2、he Estimate1.908(a).824.736625.883a Predictors: (Constant), X6, X3, X2, X4, X5ANOVA(b)Model Sum of SquaresdfMean SquareFSig.1Regression18304858.41653660971.6839.346.002(a)Residual3917298.52210391729.852 Total22222156.93715 a Predictors: (Constant), X6, X3, X2, X4, X5b Dependent Variable: YCoefficien
3、ts(a)Model Unstandardized CoefficientsStandardized CoefficientstSig.95% Confidence Interval for BBStd. ErrorBetaLower BoundUpper Bound1(Constant)5922.8272504.315 2.365.040342.86511502.790X24.8642.507.6771.940.081-.72310.451X32.374.842.7822.818.018.4974.251X4-817.901187.279-1.156-4.367.001-1235.184-4
4、00.618X514.539147.078.050.099.923-313.171342.248X6-846.867291.634-.899-2.904.016-1496.667-197.067a Dependent Variable: Y(2)用后退法选择自变量。经过后退法=0.1，得到自变量为x2,x3,x4,x6Model SummaryModelRR SquareAdjusted R SquareStd. Error of the Estimate1.908(a).824.736625.8832.907(b).824.759597.048a Predictors: (Constant)
5、, X6, X3, X2, X4, X5b Predictors: (Constant), X6, X3, X2, X4ANOVA(c)Model Sum of SquaresdfMean SquareFSig.1Regression18304858.41653660971.6839.346.002(a)Residual3917298.52210391729.852 Total22222156.93715 2Regression18301030.67544575257.66912.835.000(b)Residual3921126.26211356466.024 Total22222156.9
6、3715 a Predictors: (Constant), X6, X3, X2, X4, X5b Predictors: (Constant), X6, X3, X2, X4c Dependent Variable: YCoefficients(a)Model Unstandardized CoefficientsStandardized CoefficientstSig.95% Confidence Interval for B BStd. ErrorBeta Lower BoundUpper Bound1(Constant)5922.8272504.315 2.365.040342.8
7、6511502.790 X24.8642.507.6771.940.081-.72310.451 X32.374.842.7822.818.018.4974.251 X4-817.901187.279-1.156-4.367.001-1235.184-400.618 X514.539147.078.050.099.923-313.171342.248 X6-846.867291.634-.899-2.904.016-1496.667-197.0672(Constant)6007.3202245.481 2.675.0221065.05110949.590 X25.0681.360.7063.7
8、27.0032.0758.061 X32.308.486.7604.750.0011.2383.377 X4-824.261167.776-1.165-4.913.000-1193.535-454.988 X6-862.699232.489-.916-3.711.003-1374.403-350.995a Dependent Variable: YExcluded Variables(b)Model Beta IntSig.Partial CorrelationCollinearity StatisticsTolerance2X5.050(a).099.923.031.068a Predict
9、ors in the Model: (Constant), X6, X3, X2, X4b Dependent Variable: Y(3)用逐步回归法选择自变量。由软件得自变量为x3,x4,x5Coefficients(a)Model Unstandardized CoefficientsStandardized CoefficientstSig.95% Confidence Interval for BBStd. ErrorBetaLower BoundUpper Bound1(Constant)5161.2591142.744 4.517.0002710.3187612.201X31.5
10、11.704.4982.146.050.0013.0212(Constant)472.2982150.138 .220.830-4172.7935117.388 X33.188.9131.0503.492.0041.2165.160X5212.32586.643.7372.451.02925.144399.5053(Constant)1412.8071865.912 .757.464-2652.6675478.281 X33.440.7821.1334.398.0011.7365.144X5348.72992.2201.2103.782.003147.800549.659X4-415.1361
11、69.163-.587-2.454.030-783.712-46.561a Dependent Variable: YModel SummaryModelRR SquareAdjusted R SquareStd. Error of the Estimate1.498(a).248.1941092.8322.697(b).485.406937.9503.811(c).657.572796.609a Predictors: (Constant), X3b Predictors: (Constant), X3, X5c Predictors: (Constant), X3, X5, X4ANOVA
12、(d)Model Sum of SquaresdfMean SquareFSig.1Regression5502210.09015502210.0904.607.050(a)Residual16719946.847141194281.918 Total22222156.93715 2Regression10785395.10825392697.5546.130.013(b)Residual11436761.83013879750.910 Total22222156.93715 3Regression14607124.51934869041.5067.673.004(c)Residual7615
13、032.41812634586.035 Total22222156.93715 a Predictors: (Constant), X3b Predictors: (Constant), X3, X5c Predictors: (Constant), X3, X5, X4d Dependent Variable: Y(4)分析差异。后退法是将全部的变量先建立一个回归方程，然后逐步将回归系数F检验值最小的剔除，最终它只剔除了x5，它只是考虑了所有变量时x5的显著性最差，却没有考虑他在其他情况下可能是显著地，而且后退法不可以再把x5引进所以导致最终结果有误差；逐步回归法吸收了前进法和后退法的优点，变量有进有出，当引
