《异方差、自相关检验》由会员分享，可在线阅读，更多相关《异方差、自相关检验（9页珍藏版）》请在金锄头文库上搜索。
1、计量经济学作业It 、.几一、异方差(1)Dependent Variable: YMethod: Least SquaresDate: 09/29/11 Time: 22:09Sample: 1 29Included observations: 29VariableCoefficientStd. Error t-StatisticProb.C58.3179149.04935 1.1889640.2448X0.7955700.018373 43.301930.0000R-squared0.985805Mean dependent var2111.931Adjusted R-squared0.9
2、85279S.D. dependent var555.5470S.E. of regression67.40436Akaike info criterion11.32577Sum squared resid122670.4Schwarz criterion11.42006Log likelihood-162.2236F-statistic1875.057Durbin-Watson stat1.893970Prob(F-statistic)0.000000（1）戈徳菲尔德匡特检验：简单步骤如下：1、 先排列2、分成两组 1-11，19-29，做回归检验，得残差平方和3、 得 F ，查表比较。De
3、pendent Variable: YMethod: Least SquaresDate: 09/29/11 Time: 19:38Sample: 1 11Included observations: 11VariableCoefficientStd. Errort-StatisticProb.C55.8484060.155270.9284040.3774X0.8027690.02158637.189300.0000R-squared0.993535Mean dependent var2203.182Adjusted R-squared0.992816S.D. dependent var660
4、.2351S.E. of regression55.95928Akaike info criterion11.05009Sum squared resid28182.97Schwarz criterion11.12244Log likelihood-58.77550F-statistic1383.044Durbin-Watson stat 1.657950 Prob(F-statistic) 0.000000第一组：Sum squared resid (残差平方和)=28182.97Dependent Variable: YMethod: Least SquaresDate: 09/29/11
5、 Time: 19:39Sample: 19 29Included observations: 11VariableCoefficientStd. Errort-StatisticProb.C92.4461596.012930.9628510.3608X0.7822810.03536922.117980.0000R-squared0.981935Mean dependent var2141.455Adjusted R-squared0.979928S.D. dependent var590.5276S.E. of regression83.66352Akaike info criterion1
6、1.85445Sum squared resid62996.26Schwarz criterion11.92679Log likelihood-63.19947F-statistic489.2051Durbin-Watson stat1.770865Prob(F-statistic)0.000000第二组： Sum squared resid (残差平方和) =62996.26F=62996.26/28182.97=2.23526，给定显著性水平 a=0.05 查F分布临界值表可得临界值F0.05 (11,11) =2.85, 所以统计量F F0.05 (11,11)，支出模型不存在异方差。(
7、 2)利用加权最小二乘法估计如下：权重先用1 /x,检验结果并不是很理想，因为R方较小，拟合比较差，所以陆续用了 1/X7, X(-1/2), 效果还好。Dependent Variable: YMethod: Least SquaresDate: 09/29/11 Time: 19:47Sample: 1 29Included observations: 29Weight ing series: 1/XA2VariableCoefficientStd. Errort-StatisticProb.C103.774768.932651.5054500.1438X0.7769270.0308542
8、5.180600.0000Weighted StatisticsR-squared0.978844Mean dependent var1922.584Adjusted R-squared0.978061S.D. dependent var431.8719S.E. of regression63.96817Akaike info criterion11.22112Sum squared resid110482.0Schwarz criterion11.31542Log likelihood-160.7062F-statistic634.0624Durbin-Watson stat1.841591
9、Prob(F-statistic)0.000000Unweighted StatisticsR-squared0.985240Mean dependent var2111.931Adjusted R-squared0.984693S.D. dependent var555.5470S.E. of regression68.73310Sum squared resid127554.5Durbin-Watson stat1.921310利用加权最小二乘法估计模型如下：Y=103.7747+0.776927XDependent Variable: YMethod: Least SquaresDate
10、: 09/29/11 Time: 22:29Sample: 1 29Included observations: 29Weighti ng series: XA(-1/2)VariableCoefficientStd. Errort-StatisticProb.C73.8263353.09180 1.3905410.1757X0.7895620.021139 37.351170.0000Weighted StatisticsR-squared0.932667Mean dependent var2053.653Adjusted R-squared0.930173S.D. dependent va
11、r251.3231S.E. of regression66.41175Akaike info criterion11.29610Sum squared resid119084.1Schwarz criterion11.39039Log likelihood-161.7934F-statistic1395.110Durbin-Watson stat1.882707Prob(F-statistic)0.000000Unweighted StatisticsR-squared0.985749Mean dependent var2111.931Adjusted R-squared0.985221S.D
12、. dependent var555.5470S.E. of regressionDurbin-Watson stat67.53770 Sum squared resid123156.21.906531利用加权最小二乘法估计模型如下Y=73.82633+0.789562X二、自相关1)Dependent Variable: YMethod: Least SquaresDate: 09/29/11 Time: 19:49Sample: 1960 2001Included observations: 42VariableCoefficientStd. Error t-StatisticProb.C
13、-3028.503655.4104 -4.6207730.0000GDP0.6974810.019059 36.595540.0000R-squared0.970998Mean dependent var10765.23Adjusted R-squared0.970273S.D. dependent var20154.12S.E. of regression3474.857Akaike info criterion19.19094Sum squared resid4.83E+08Schwarz criterion19.27369Log likelihood-401.0098F-statisti
14、c1339.234Durbin-Watson stat0.178429Prob(F-statistic)0.000000y=-3028.503+0.697481gdp(4) (4) p=DW/2=0.9107855作两次广义差分：Dependent Variable: YYMethod: Least SquaresDate: 09/29/11 Time: 19:50Sample (adjusted): 1961 2001Included observations: 41 after adjustmentsVariableCoefficientStd. Error t-StatisticProb.C-421.6400280.8793-1.5011430.1414GG0.7795260.04299018.132730.0000R-squared0.893963Mean dependent var2620.801Adjus
