7、Std. Errort-StatisticProb. X50.0025310.0012721.9891890.0582X4155.432328.258835.5003090.0000X30.0002040.0007220.2828560.7797X2-1.4702420.596051-2.4666360.0212X12.4362740.8837002.7569010.0110C-38.32851692.8780-0.0553180.9563R-squared0.830540Mean dependent var4829.185Adjusted R-squared0.795235S.D. depe 。

8、ndent var2431.177S.E. of regression1100.130Akaike info criterion17.02110Sum squared resid29046879Schwarz criterion17.30134Log likelihood-249.3165F-statistic23.52519Durbin-Watson stat1.767723Prob(F-statistic)0.000000从表中发现在0.1时 ， 虽然模型拟合效果较好 ， F检验显著 ， 但X3的t检验不显著 ， X2系数的符号为负 ， 与经济意义不服 ， 则模型存在多重共线 ， 故采用逐步回归法进行修正 。
2、逐步回 。

9、归对X1、X2、X3、X4、X5单个回归后发现X4的拟合程度最好 ， 结果为：VariableCoefficientStd. Errort-StatisticProb. X4167.107928.121365.9423840.0000C1011.500702.01771.4408470.1603R-squared0.549073Mean dependent var4799.998Adjusted R-squared0.533524S.D. dependent var2395.831S.E. of regression1636.331Akaike info criterion17.70064Sum。

10、squared resid77649780Schwarz criterion17.79316Log likelihood-272.3599F-statistic35.31192Durbin-Watson stat1.629401Prob(F-statistic)0.000002再将其余解释变量逐个加入后 ， 得结果为：VariableCoefficientStd. Errort-StatisticProb. X4192.633125.883087.4424340.0000X50.0023430.0007922.9569300.0064C-149.6733633.1428-0.2363970.814 。

11、9R-squared0.751001Mean dependent var4829.185Adjusted R-squared0.732557S.D. dependent var2431.177S.E. of regression1257.281Akaike info criterion17.20593Sum squared resid42680430Schwarz criterion17.34605Log likelihood-255.0890F-statistic40.71708Durbin-Watson stat1.739272Prob(F-statistic)0.000000继续加入 ， 得 。

12、结果：VariableCoefficientStd. Errort-StatisticProb. X50.0023240.0007653.0388380.0054X4161.285030.861175.2261470.0000X30.0011940.0006911.7295980.0956C314.9939667.46460.4719260.6409R-squared0.776694Mean dependent var4829.185Adjusted R-squared0.750928S.D. dependent var2431.177S.E. of regression1213.330Aka 。

13、ike info criterion17.16369Sum squared resid38276420Schwarz criterion17.35052Log likelihood-253.4554F-statistic30.14407Durbin-Watson stat2.073796Prob(F-statistic)0.000000以及：VariableCoefficientStd. Errort-StatisticProb. X50.0011860.0012610.9403540.3560X4163.793730.775615.3221930.0000X30.0010100.000706 。

14、1.4304820.1650X10.4359340.3851521.1318500.2684C-102.6452759.5370-0.1351420.8936R-squared0.787579Mean dependent var4829.185Adjusted R-squared0.753592S.D. dependent var2431.177S.E. of regression1206.824Akaike info criterion17.18038Sum squared resid36410617Schwarz criterion17.41392Log likelihood-252.70 。

15、58F-statistic23.17274Durbin-Watson stat1.988861Prob(F-statistic)0.000000可以发现 ， 在加入了四个解释变量后 ， 模型的回归效果不如只有解释变量X3、X4、X5时效果好 ， 所以通过逐步回归后模型为Y=314.9939+0.001194 X3+161.285 X4+0.002324 X53、异方差检验对Y=314.9939+0.001194 X3+161.285 X4+0.002324 X5进行检验得：White Heteroskedasticity Test:F-statistic1.291175Probability0.30113 。

16、6Obs*R-squared11.02501Probability0.273998当0.1时Obs*R-squared 11.0250120。
1（3）6.251 ， 所以应拒绝原假设 ， 认为模型存在异方差 。
在此 ， 选择WLS法进行修正 ， 权重W1/e ， 修正结果为：Weighting series: W1VariableCoefficientStd. Errort-StatisticProb. X50.0020970.00016212.980280.0000X488.539645.26392416.820080.0000X30.0003170.0002321.3634820.1844C1930.2297 。

17、2.5000326.623840.0000Weighted StatisticsR-squared0.999843Mean dependent var3521.571Adjusted R-squared0.999825S.D. dependent var23095.27S.E. of regression305.7747Akaike info criterion14.40714Sum squared resid2430953.Schwarz criterion14.59397Log likelihood-212.1071F-statistic602.7000Durbin-Watson stat 。

18、0.146268Prob(F-statistic)0.000000Unweighted StatisticsR-squared0.598751Mean dependent var4829.185Adjusted R-squared0.552453S.D. dependent var2431.177S.E. of regression1626.431Sum squared resid68777250Durbin-Watson stat1.608504进行异方差检验为：White Heteroskedasticity Test:F-statistic0.567448Probability0.807 。

19、602Obs*R-squared6.102317Probability0.729636此时Obs*R-squared6.10231720。
1（3）6.251 ， 则接受原假设 ， 认为模型不存在异方差 ， 且拟合程度高 ， 然而X3的t检验却稍不显著 ， 但此模型在所有模型中效果最好 ， 且从本文的目的而言 ， 也是可以选择此模型的 。
而此模型中Durbin-Watson stat 0.146268落在了正自相关区域 ， 于是进行自相关修正 。
4、自相关修正在此 ， 采用了迭代法 ， 一次迭代后结果为：Dependent Variable: YMethod: Least SquaresDate: 12/19/04 Time: 18:0 。

20、7Sample(adjusted): 2 31Included observations: 28Excluded observations: 2 after adjusting endpointsConvergence achieved after 8 iterationsVariableCoefficientStd. Errort-StatisticProb. X50.0023250.0008232.8252060.0096X4129.925831.461914.1296240.0004X30.0034610.0014812.3377280.0285C679.4463639.09711.06 。

稿源：(未知)

【傻大方】网址：/a/2021/0822/0023896414.html

标题：地区|地区人均收入影响因素的计量分析( 二 )