1、第六章 多重共线性行情形及处理习题 Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. Collinearity Statistics B Std. Error Beta Tolerance VIF 1 (Constant) 1348.338 2211.463 .610 .552 农业x1 -.641 .167 -1.125 -3.840 .002 .003 319.484 工业x2 -.317 .204 -1
2、306 -1.551 .143 .000 2.637E3 建筑业x3 -.413 .548 -.270 -.752 .464 .002 479.288 人口x4 -.002 .024 -.007 -.087 .932 .037 27.177 最终消费x5 .671 .128 3.706 5.241 .000 .001 1.861E3 受灾面积x6 -.008 .008 -.020 -.928 .369 .574 1.743 a. Dependent Variable: 财政收入y 书
3、 Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. Collinearity Statistics B Std. Error Beta Tolerance VIF 1 (Constant) 1348.338 2211.463 .610 .552 农业x1 -.641 .167 -1.125 -3.840 .002 .003 319.484 工业x2 -.317 .204 -1.306 -1.551 .143
4、000 2.637E3 建筑业x3 -.413 .548 -.270 -.752 .464 .002 479.288 人口x4 -.002 .024 -.007 -.087 .932 .037 27.177 最终消费x5 .671 .128 3.706 5.241 .000 .001 1.861E3 受灾面积x6 -.008 .008 -.020 -.928 .369 .574 1.743 a. Dependent Variable: 财政收入y 6.6解:将5.9数据录入spss ,在线性回归
5、对话框里选择Statistics选择Collinear Diagnostic 做共线性判断然后做回归得到输出结果如下表: 从输出结果来看的扩大因子最大,=2637,远大于10,所以存在严重共线线性。因此剔除,继续建立变量y,关于的回归方程,有关结果输出如下: Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. Collinearity Statistics B Std. Error Bet
6、a Tolerance VIF 1 (Constant) -1252.832 1507.836 -.831 .419 农业x1 -.735 .163 -1.291 -4.524 .000 .004 276.969 建筑业x3 -.923 .459 -.604 -2.012 .063 .003 306.617 人口x4 .026 .017 .093 1.591 .132 .086 11.605 最终消费x5 .510 .078 2.815 6.527 .000 .002 632.896 受灾面积x6
7、011 .008 -.028 -1.274 .222 .608 1.645 a. Dependent Variable: 财政收入y Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. Collinearity Statistics B Std. Error Beta Tolerance VIF 1 (Constant) -2715.046 2829.351 -.960 .352 农业x1 -
8、047 .235 -.083 -.202 .843 .006 160.513 建筑业x3 1.463 .526 .957 2.781 .013 .009 111.949 人口x4 .036 .031 .128 1.160 .263 .087 11.507 受灾面积x6 .003 .015 .008 .206 .839 .649 1.540 a. Dependent Variable: 财政收入y 从输出结果看,的方差扩大因子=632.896依然远大于10,所以存在多重共线性性剔除,继续建立变量y关于做
9、回归,输出如下: Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. Collinearity Statistics B Std. Error Beta Tolerance VIF 1 (Constant) -2296.322 1869.710 -1.228 .236 建筑业x3 1.359 .097 .889 14.036 .000 .249 4.018
10、 人口x4 .031 .019 .111 1.649 .117 .222 4.509 受灾面积x6 .004 .014 .010 .256 .801 .673 1.485 a. Dependent Variable: 财政收入y 从输出结果看,的方差扩大因子=160.513依然远大于10,所以存在多重共线性性剔除,继续建立变量y关于做回归,输出如下: 从输出结果看到,3个方差扩大因子都小于10,回归系数也有合理的解释,说明回归模型不存在强多重共线性。可以作为最终回归
11、模型。回归方程为: 标准化的回归方程为: Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. Collinearity Statistics B Std. Error Beta Tolerance VIF 1 (Constant) -1252.832 1507.836 -.831 .419 农业x1 -.735 .163 -1.291 -4.524 .000 .004 276.969 建筑业x3 -.923 .459 -.604 -2.012 .063 .003 306.617 人口x4 .026 .017 .093 1.591 .132 .086 11.605 最终消费x5 .510 .078 2.815 6.527 .000 .002 632.896 受灾面积x6 -.011 .008 -.028 -1.274 .222 .608 1.645 a. Dependent Variable: 财政收入y






