【正文】 var. of regressionAkaike info criterionSum squared residSchwarz criterionLog likelihoodHannanQuinn criter.FstatisticDurbinWatson statProb(Fstatistic)①得到模型的方程為：Y= X6+②評價：1） ，數(shù)據(jù)相當(dāng)大，可以認(rèn)為擬合程度很好。 X6 ，小于t（12）=，所以系數(shù)是不顯著的。但當(dāng)α=，t（24）=，LNCPI的系數(shù)不顯著，可能存在多重共線性。LNGDPLNGDP， LNCPI之間的相關(guān)系數(shù)很高，證實確實存在多重共線性。（4）建議：如果僅僅是作預(yù)測，可以不在意這種多重共線性，但如果是進行結(jié)構(gòu)分析，還是應(yīng)該引起注意的。說明在α=,水平下，回歸方程回歸方程整體上是顯著的。由此可得知，該方程可能存在多重共線性。（3）做輔助回歸被解釋變量可決系數(shù)方差擴大因子CZZC353GDP90SSZE468方差擴大因子均大于10，存在嚴(yán)重多重共線性。（4）解決方式：分別作出財政收入與財政支出、國內(nèi)生產(chǎn)總值、稅收總額之間的一元回歸。（2）存在異方差，估計參數(shù)的方法：①可以對模型進行變換②使用加權(quán)最小二乘法進行計算，得出模型方程，并對其進行相關(guān)檢驗③對模型進行對數(shù)變換，進行分析（3）評價：，隨機擾動項之間不存在異方差。(1)由Eviews軟件分析得：Dependent Variable: YMethod: Least SquaresDate: 12/24/15 Time: 16:00Sample: 1 31Included observations: 31VariableCoefficientStd. ErrortStatisticProb.XCRsquaredMean dependent varAdjusted Rsquared. dependent var. of regressionAkaike info criterionSum squared resid12220196Schwarz criterionLog likelihoodHannanQuinn criter.FstatisticDurbinWatson statProb(Fstatistic)由上表可知，2007年我國農(nóng)村居民家庭人均消費支出（x）對人均純收入（y）的模型為：Y=+（2）①由圖形法檢驗由上圖可知，模型可能存在異方差。（3）1）采用WLS法估計過程中，①用權(quán)數(shù)w1=1/X,建立回歸得：Dependent Variable: YMethod: Least SquaresDate: 12/24/15 Time: 16:29Sample: 1 31Included observations: 31Weighting series: W1VariableCoefficientStd. ErrortStatisticProb.XCWeighted StatisticsRsquaredMean dependent varAdjusted Rsquared. dependent var. of regressionAkaike info criterionSum squared resid8352726.Schwarz criterionLog likelihoodHannanQuinn criter.FstatisticDurbinWatson statProb(Fstatistic)Unweighted StatisticsRsquaredMean dependent varAdjusted Rsquared. dependent var. of regressionSum squared resid14484289DurbinWatson stat對此模型進行White檢驗得：Heteroskedasticity Test: WhiteFstatisticProb. F(2,28)Obs*RsquaredProb. ChiSquare(2)Scaled explained SSProb. ChiSquare(2)Test Equation:Dependent Variable: WGT_RESID^2Method: Least SquaresDate: 12/24/15 Time: 16:34Sample: 1 31Included observations: 31Collinear test regressors dropped from specificationVariableCoefficientStd. ErrortStatisticProb.C1045682.WGT^21173622.X*WGT^2RsquaredMean dependent varAdjusted Rsquared. dependent var. of regressionAkaike info criterionSum squared resid+13Schwarz criterionLog likelihoodHannanQuinn criter.FstatisticDurbinWatson statProb(Fstatistic)從上可知，nR2=，比較計算的統(tǒng)計量的臨界值，因為nR2=（2）=，所以接受原假設(shè)，該模型消除了異方差。估計結(jié)果為： Y= t=（）（）R2= F= DW=③用權(quán)數(shù)w3=1/sqr（x）,用回歸分析得：Dependent Variable: YMethod: Least SquaresDate: 12/24/15 Time: 16:49Sample: 1 31Included observations: 31Weighting series: W3VariableCoefficientStd. ErrortStatisticProb.XCWeighted StatisticsRsquaredMean dependent varAdjusted Rsquared. dependent var. of regressionAkaike info criterionSum squared resid9990985.Schwarz criterionLog likelihoodHannanQuinn criter.FstatisticDurbinWatson statProb(Fstatistic)Unweighted StatisticsRsquaredMean dependent varAdjusted Rsquared. dependent var. of regressionSum squared resid12717412DurbinWatson stat對此模型進行White檢驗得：Heteroskedasticity Test: WhiteFstatisticProb. F(2,28)Obs*RsquaredProb. ChiSquare(2)Scaled explained SSProb. ChiSquare(2)Test Equation:Dependent Variable: WGT_RESID^2Method: Least SquaresDate: 12/24/15 Time: 16:57Sample: 1 31Included observations: 31Collinear test regressors dropped from specificationVariableCoefficientStd. ErrortStatisticProb.C1212308.2141958.WGT^21301839.X^2*WGT^2RsquaredMean dependent varAdjusted Rsquared. dependent var. of regressionAkaike info criterionSum squared resid+13Schwarz criterionLog likelihoodHannanQuinn criter.FstatisticDurbinWatson statProb(Fstatistic)從上可知，nR2=，比較計算的統(tǒng)計量的臨界值，因為nR2=（2）=，所以接受原假設(shè)，該模型消除了異方差。②White檢驗用EViews軟件分析得：Heteroskedasticity Test: WhiteFstatisticProb. F(2,31)Obs*RsquaredProb. ChiSquare(2)Scaled explained SSProb. ChiSquare(2)Test Equation:Dependent Variable: RESID^2Method: Least SquaresDate: 12/24/15 Time: 19:56Sample: 1 34Included observations: 34VariableCoefficientStd. ErrortStatistic