应用回归分析实验三：多元线性回归

《应用回归分析实验三：多元线性回归》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、实验三：多元线性回归实验内容习题一（P64例3・1）（1）打开SPSS软件，输入数据如下（部分）:■1地区X1x2x3x4x5x6x7□北京5562110212861563809.355844夭津500581315281221715239990河北31557551097809386.624276山西29758217701570.825489内蒙古355311781028870568.425949辽宁43787221271913670.127179吉林33078711285915576.2232941黑龙江31288

2、31941865551.721764上海7109113516467551166.052122江苏45458571042795751031297浙江5523118413349331179.033622安徴3905679988634633.5福建50798621300541686.425555江西3633739851484388520597山东36999971247800580.126234■

3、河南30808429647911464.424438湖北39968169146751571.022384选择“分析”中“回归“

4、线性”,以y为应变量,以X1-X9为自变量,点击“确定”得:模型'lli•ItSig.B拆准谋差试用版1（常量）-2721.4938050.265-.338.739X11.465.151.5209.715.000x22.575.422.1676.103.000x32.005.322.1876.236.000x4.000.507.000.001.999x5.891.474.0661.881.075x6.067.019.1983.491.002x7.028.015.0451.826.083x811.40570.975.

5、005.161.874x9-160.711121.989-.032-1.317.203a.因变虽:y所以得回归方程为:y=l.465X1+2.575X2+2.005X3+0.891X5+0.67X6+0.28X7+II.405x8-160.711X9-2721.493从回国方程可以看到，xl-x9对屈民的消费支出起正影响，x9对屈民的消费性支出起负影响。（2）F检验。用SPSS软件计算出的方差分析图如下：Anova3模型平方和df均方FSig.1回归残差总计216165880.51305283.4672174711

6、64.09202924018431.1765264.173368.019.000ba.因变量:yb.预測变说:（常動x9,x8,x2,x1,x7,x3,x4tx5,x6从输出结果可知，Sig即显著性P值，由P值为0.000可知，此回归方程高度显著。t检验。通过定性分析，先剔除x4,用y与其他8个变量做冋山分析，计算结果如下图:模型非标准化系教拆淮系数tSig.B标准谋差1（常E）-2716.2715698.036-.477.638X11.465.108.52013.577.000x22.575.411.1676.2

7、65.000x32.006.302.1886.652.000x5.891.461.0661.931.067x6.067.013.1985.065.000x7.028.014.0451.916.069x811.36051.253.005.222.827x9-160.749112.288-.032-1.432.167a因变就y剔除x4Z后，仍然有不显著的自变量，此时最大的P值为p8=0.827，因此进一步剔除x8,用y与其余6个变量作回归，回归系数表如下图:模型非标淮化系数标淮系数tSig.B标淮i5?差1（常E）・1

8、459.222539.332-2.706.013X11.455.096.51615.128.000x22.539.370.1646.872.000x31.990.287.1866.935.000x5.898.450.0671.996.059x6.068.012.2025.759.000x7.027.014.0451.946.065x9-160.194109.807-.032-1.459.159a.因变虽:yT检验中，依次剔除P值最大的自变量，直到最后所有的自变量在显著性水平为0.05时都显著。习题二（卩93・例4・

9、3）（1）打开spss软件,输入数据如F图:1X丄8777.00IyJ264.009210.00105.009954.0090.0010508.00131.0010979.00122.0011912.00107.0012747.00406.0013499.00503.0014269.00431.0015522.00588.0016730.00898.00MM