計量經(jīng)濟(jì)學(xué)虛擬變量實驗報告計量經(jīng)濟(jì)學(xué)課程實驗報告

1、第69頁 共69頁計量經(jīng)濟(jì)學(xué)虛擬變量實驗報告 計量經(jīng)濟(jì)學(xué)課程實驗報告 計量經(jīng)濟(jì)學(xué)課程實驗報告實驗序號2實驗名稱Eviews的異方差檢驗與校正實驗組別12模擬角色實驗地點2教602指導(dǎo)老師劉冬萍實驗日期11月29日實驗時間16:0517:45一、實驗?zāi)康募耙髮W(xué)會使用計量學(xué)分析p p 軟件Eviews的異方差檢驗與校正功能。二、實驗環(huán)境2教602,經(jīng)管學(xué)院電腦實驗室三、實驗內(nèi)容與步驟 ?DATA Y _SORT _SCAT _ Y根據(jù)相關(guān)圖隨著_的增大Y的取值范圍不斷增大,所以方程存在異方差.（1）WHITE 檢驗建立回歸模型 LS Y C _ Dependent Variable: YMeth

2、od: Least SquaresDate: 11/22/12 Time: 17:06Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.C_R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson stat

3、Prob(F-statistic)進(jìn)行WHITE 檢驗White Heteroskedasticity Test:F-statisticProbabilityObs_R-squaredProbabilityTest Equation:Dependent Variable: RESID2Method: Least SquaresDate: 11/22/12 Time: 17:07Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.C_2R-squaredMean dependent varAd

4、justed R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)Nr2=8.413677 因為檢驗的P=0.014893小于0.05,所以存在異方差.(2) PARK檢驗LS Y C _Dependent Variable: YMethod: Least SquaresDate: 11/22/12 Time: 17:13Sle: 1

5、 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.C_R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)GENR E2=LOG(RESID2)GENR LN_=LOG(

6、_)LS LNE2 C LN_Dependent Variable: LNE2Method: Least SquaresDate: 11/22/12 Time: 17:16Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.CLN_R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criteri

7、onLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)由上圖可看出P分別為0.0033 ,0.0048,0.004754都是小概率事件,所以方程是顯著的,表明隨機誤差項的方差隨著解釋變量的取值不同而不斷變化,即存在異方差性.（3）GLEISER檢驗LS Y C _GENR E=ABS(RESID)LS E C _1Dependent Variable: E1Method: Least SquaresDate: 11/28/12 Time: 13:14Sle: 1 20Included observations: 20Vari

8、ableCoefficientStd.Errort-StatisticProb.C_1R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)R2=0.473046 F=16.15859 P=eq oac(,2)GENR _2=_-2LS E C _2Dependen

9、t Variable: EMethod: Least SquaresDate: 11/28/12 Time: 13:27Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.C_2R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statistic

10、Durbin-Watson statProb(F-statistic)|e2|=1.665123-657.9505_-2R2=0.173874 F=3.788442 P=eq oac(,3)GENR _3=_2LS E C _3Dependent Variable: EMethod: Least SquaresDate: 11/28/12 Time: 13:32Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.C_3R-squaredMean dependent varAdjusted R

11、-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)|e3|=0.580535+0.000113_42eq oac(,4)GENR _4=_-0,5LS E C _4Dependent Variable: EMethod: Least SquaresDate: 11/28/12 Time: 13:36Sle: 1 20Included

12、observations: 20VariableCoefficientStd.Errort-StatisticProb.C_4R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)R2=0.350914 F=9.731299 P=eq oac(,5)GENR _5=

13、_-1LS E C _5Dependent Variable: EMethod: Least SquaresDate: 11/28/12 Time: 13:45Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.C_5R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog l

14、ikelihoodF-statisticDurbin-Watson statProb(F-statistic)|e5|=2.265778-45.87625_-1由以上的五個方程表明，利潤函數(shù)存在異方差性（只要取顯著水平a大于0.067388）(1)利用最小二乘法估計模型LS Y C _Dependent Variable: YMethod: Least SquaresDate: 11/28/12 Time: 12:40Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.C_R-squared

15、Mean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)(0.0014)T=(1.212130) (3.772393 )(2)生成權(quán)數(shù)變量：根據(jù)帕克檢驗得到：Ls y c _Genr lne2=log(resid2)Genr ln_=log(_)Ls lne2 c ln_Depende

16、nt Variable: LNE2Method: Least SquaresDate: 11/28/12 Time: 12:56Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.CLN_R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-stat

17、isticDurbin-Watson statProb(F-statistic)進(jìn)行戈里瑟檢驗LS Y C _GENR E=ABS(RESID)LS E C _1Dependent Variable: E1Method: Least SquaresDate: 11/28/12 Time: 13:14Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.C_1R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of

18、regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)R2=0.473046 F=16.15859 P=eq oac(,2)GENR _2=_-2LS E C _2Dependent Variable: EMethod: Least SquaresDate: 11/28/12 Time: 13:27Sle: 1 20Included observations: 20VariableCoefficien

19、tStd.Errort-StatisticProb.C_2R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)|e2|=1.665123-657.9505_-2R2=0.173874 F=3.788442 P=eq oac(,3)GENR _3=_2LS E C

20、_3Dependent Variable: EMethod: Least SquaresDate: 11/28/12 Time: 13:32Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.C_3R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF

21、-statisticDurbin-Watson statProb(F-statistic)|e3|=0.580535+0.000113_42eq oac(,4)GENR _4=_-0,5LS E C _4Dependent Variable: EMethod: Least SquaresDate: 11/28/12 Time: 13:36Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.C_4R-squaredMean dependent varAdjusted R-squaredS.D.

22、dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)R2=0.350914 F=9.731299 P=eq oac(,5)GENR _5=_-1LS E C _5Dependent Variable: EMethod: Least SquaresDate: 11/28/12 Time: 13:45Sle: 1 20Included observations:

23、20VariableCoefficientStd.Errort-StatisticProb.C_5R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)|e5|=2.265778-45.87625_-1R2=0.281461 F=7.050824 P=由上可得在戈里

24、瑟檢驗里最顯著的是：|e3|=0.580535+0.000113_42R2=0.498972 F=17.92617 P=GENR W2=_2另外取：GENR W3=1/ABS(RESID)GENR W4=1/RESID2（3）利用最小二乘法估計模型：模型一LS(W=W1) Y C _ Dependent Variable: YMethod: Least SquaresDate: 11/28/12 Time: 14:00Sle: 1 20Included observations: 20Weighting series: W1VariableCoefficientStd.Errort-Stati

25、sticProb.C_Weighted StatisticsR-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)Unweighted StatisticsR-squaredMean dependent varAdjusted R-squaredS.D.depend

26、ent varS.E.of regressionSum squared residDurbin-Watson stat懷特檢驗的結(jié)果是White Heteroskedasticity Test:F-statisticProbabilityObs_R-squaredProbabilityTest Equation:Dependent Variable: STD_RESID2Method: Least SquaresDate: 11/28/12 Time: 14:36Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-St

27、atisticProb.C_2R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)(0.318225) (6.100161)模型二LS(W=W2) Y C _Dependent Variable: YMethod: Least SquaresDate: 11/28

28、/12 Time: 14:12Sle: 1 20Included observations: 20Weighting series: W2VariableCoefficientStd.Errort-StatisticProb.C_Weighted StatisticsR-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-

29、Watson statProb(F-statistic)Unweighted StatisticsR-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionSum squared residDurbin-Watson stat進(jìn)行懷特檢驗的結(jié)果是White Heteroskedasticity Test:F-statisticProbabilityObs_R-squaredProbabilityTest Equation:Dependent Variable: STD_RESID2Method:

30、 Least SquaresDate: 11/28/12 Time: 14:39Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.C_2R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statPr

31、ob(F-statistic)(3.255974) (0.022701)模型三LS(W=W3) Y C _Dependent Variable: YMethod: Least SquaresDate: 11/28/12 Time: 14:19Sle: 1 20Included observations: 20Weighting series: W3VariableCoefficientStd.Errort-StatisticProb.C_Weighted StatisticsR-squaredMean dependent varAdjusted R-squaredS.D.dependent v

32、arS.E.of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)Unweighted StatisticsR-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionSum squared residDurbin-Watson stat進(jìn)行懷特檢驗得White Heteroskedasticity

33、 Test:F-statisticProbabilityObs_R-squaredProbabilityTest Equation:Dependent Variable: STD_RESID2Method: Least SquaresDate: 11/28/12 Time: 14:40Sle: 1 20Included observations: 20VariableCoefficientStd.Errort-StatisticProb.C_2R-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regress

34、ionAkaike info criterionSum squared residSchwarz criterionLog likelihoodF-statisticDurbin-Watson statProb(F-statistic)(0.208266) (0.005388)模型四 LS(W=W4) Y C _Dependent Variable: YMethod: Least SquaresDate: 11/28/12 Time: 14:24Sle: 1 20Included observations: 20Weighting series: W4VariableCoefficientStd.Errort-StatisticProb.C_Weighted StatisticsR-squaredMean dependent varAdjusted R-squaredS.D.dependent varS.E.of regressionAkaike info criter