計(jì)量經(jīng)濟(jì)學(xué)課程論文-用計(jì)量方法驗(yàn)證我國(guó)私人汽車消費(fèi)影響因素分析(文件)
【正文】 b(Fstatistic)再做2004到2010年的回歸得：表 7 異方差的GQ檢驗(yàn)回歸結(jié)果2：Dependent Variable: YMethod: Least SquaresDate: 12/26/11 Time: 11:29Sample: 2004 2010Included observations: 7VariableCoefficientStd. ErrortStatisticProb. CX1X2Rsquared Mean dependent varAdjusted Rsquared . dependent var. of regression Akaike info criterionSum squared resid Schwarz criterionLog likelihood FstatisticDurbinWatson stat Prob(Fstatistic)得到F統(tǒng)計(jì)量F = ，分子分母自由度均為5（72）查臨界值F(5,5) = ，說(shuō)明存在異方差.（2）White檢驗(yàn)：表 8 White檢驗(yàn)結(jié)果White Heteroskedasticity Test:Fstatistic ProbabilityObs*Rsquared ProbabilityTest Equation:Dependent Variable: RESID^2Method: Least SquaresDate: 12/26/11 Time: 13:26Sample: 2003 2010Included observations: 8VariableCoefficientStd. ErrortStatisticProb. CX1X1^2X1*X2X2X2^2Rsquared Mean dependent varAdjusted Rsquared . dependent var. of regression Akaike info criterionSum squared resid+09 Schwarz criterionLog likelihood FstatisticDurbinWatson stat Prob(Fstatistic)nR^2 = 17* = ，在5%的顯著水平下，此數(shù)值小于臨界值為（5）= ，即用懷特檢驗(yàn)方法表明模型不存在異方差。接下來(lái)我們通過(guò)加權(quán)最小二乘法修正異方差，然后與之前的回歸結(jié)果進(jìn)行對(duì)比以最終確定異方差存不存在：（3）修正異方差：運(yùn)用加權(quán)最小二乘法,分別選用權(quán)數(shù)= 1/x1, = 1/x1^2, = sqr(x1) = 1/x2, = 1/x2^2, = sqr(x2), = 1/(x1*x2)，= 1/sqr(x1*x2)發(fā)現(xiàn)用=1/sqr((x2)擬合效果是里面最好的，估計(jì)結(jié)果為表 9 用權(quán)數(shù)的結(jié)果Dependent Variable: YMethod: Least SquaresDate: 12/26/11 Time: 16:42Sample: 1994 2010Included observations: 17Weighting series: 1/SQR(X2)VariableCoefficientStd. ErrortStatisticProb. CX1X2Weighted StatisticsRsquared Mean dependent varAdjusted Rsquared . dependent var. of regression Akaike info criterionSum squared resid Schwarz criterionLog likelihood FstatisticDurbinWatson stat Prob(Fstatistic)Unweighted StatisticsRsquared Mean dependent varAdjusted Rsquared . dependent var. of regression Sum squared residDurbinWatson stat我們發(fā)現(xiàn)經(jīng)修正后的回歸結(jié)果，x1的t值、p值還沒有修正之前的回歸結(jié)果好。（七）自相關(guān)性檢驗(yàn)（1） 殘差圖：殘差的變動(dòng)連續(xù)為正和連續(xù)為負(fù)，表明殘差相存在一階正自相關(guān)。為了驗(yàn)證上面廣義差分結(jié)果，我們同時(shí)做了科克倫奧克特迭代法，結(jié)果如下：表 12 科克倫奧克特迭代結(jié)果Dependent Variable: YMethod: Least SquaresDate: 12/27/11 Time: 15:11Sample(adjusted): 1995 2010Included observations: 16 after adjusting endpointsConvergence achieved after 8 iterationsVariableCoefficientStd. ErrortStatisticProb. C
畢業(yè)設(shè)計(jì)相關(guān)推薦
鄂ICP備17016276號(hào)-1