進(jìn)口總額影響因素分析及預(yù)測(編輯修改稿)
2024-07-19 19:45 本頁面
【文章內(nèi)容簡介】 11EX21EX31EX41EX51E大小FstatisticF臨界值是否存在異方差存在不存在存在存在存在由上表可知，存在異方差。對新序列處理自相關(guān)：Dependent Variable: Y1EMethod: Least SquaresDate: 12/04/13 Time: 21:41Sample(adjusted): 1979 2012Included observations: 34 after adjusting endpointsConvergence achieved after 8 iterationsVariableCoefficientStd. ErrortStatisticProb. C1EX11EX21EX31EX41EX51EAR(1)Rsquared Mean dependent varAdjusted Rsquared . dependent var. of regression Akaike info criterionSum squared resid Schwarz criterionLog likelihood DurbinWatson statInverted AR Roots .80對模型做廣義一階差分回歸，得AR（1）=，經(jīng)差分后，得到新序列，設(shè)為C1 Y1 X11 X21 X31 X41 X51 +生成新數(shù)據(jù)：C1=c1ec1e(1)*Y1=y1ey1e(1)*X11=x11ex11e(1) *X21=x21ex21e(1) *X31=x31ex31e(1) *X41=x41ex41e(1) *X51=x51ex51e(1) *補(bǔ)第一個數(shù)（1978年）：c1=c1e *(^2)^Y1=y1e*(^2)^X11=x11e*(^2)^X21=x21e*(^2)^ X31=x31e*(^2)^X41=x41e*(^2)^X51=x51e*(^2)^產(chǎn)生如下新序列：obsC1X11X21X31X41X51Y119781979198019811982198319841985198619871988198919901991199219931994199519961997199819992000200120022003200420052006200720082009201020112012下面對新序列檢驗自相關(guān)：Dependent Variable: Y1Method: Least SquaresDate: 12/04/13 Time: 22:34Sample: 1978 2012Included observations: 35VariableCoefficientStd. ErrortStatisticProb. C1X11X21X31X41X51Rsquared Mean dependent varAdjusted Rsquared . dependent var. of regression Akaike info criterionSum squared resid Schwarz criterionLog likelihood DurbinWatson stat對新序列回歸，可得DW=，查表知dL= dU=,則dU=DW=4dU=，不存在自相關(guān)。對新序列檢驗異方差：將Xi1按升序排列，i=1，2，3，4刪掉排在中間的n/4個，即35/4個，刪掉9個。將其余樣本點(diǎn)劃分為樣本容量各為13個的兩個子樣本。分別用兩個子樣本做回歸。求出大和小，并做F檢驗。結(jié)果如下表所示：X11X21X31X41X51大小FstatisticF臨界值是否存在異方差存在存在存在存在存在由上表可知，存在異方差。對新序列處理異方差：建立新序列： genr e1=abs(resid) c2=c1/e1 x12=x11/e1 x22=x21/e1 x32=x31/e1 x42=x41/e1 x52=x51/e1 y2=y1/e1名稱分別為：c2 x12 x22 x32 x42 x52 y2產(chǎn)生以下新序列：obsC2X12X22X32X42X52Y219781979198019811982198319841985198619871988198919901991199219931994199519961997199819992000200120022003200420052006200720082009201020112012下面對新序列檢驗自相關(guān)：Dependent Variable: Y2Method: Least SquaresDate: 12/04/13 Time: 23:28Sample: 1978 2012Included observations: 35VariableCoefficientStd. ErrortStatisticProb. C2X12X22X32X42X52Rsquared Mean dependent varAdjusted Rsquared . dependent var. of regression Akaike info criterionSum squared resid Schwarz criterionLog likelihood DurbinWatson stat對新序列回歸，得DW=，查表知dL= dU=,=dLDW=dU=，所以認(rèn)為存在自相關(guān)。對新序列檢驗異方差：將Xi2按升序排列，i=1，2，3，4，5刪掉排在中間的n/4個，即35/4個，刪掉9個。將其余樣本點(diǎn)劃分為樣本容量各為13個的兩個子樣本。分別用兩個子樣本做回歸。求出大和小，并做F檢驗。結(jié)果如下表所示：X12X22 X32X42X52大小FstatisticF臨界值是否存在異方差不存在不存在不存在不存在不存在由上表可知，不存在異方差。對新序列處理自相關(guān)：Dependent Variable: Y2Method: Least SquaresDate: 12/05/13 Time: 00:02Sample(adjusted): 1979 2012Included observations: 34 after adjusting endpointsConvergence achieved after 10 iterationsVariableCoefficientStd. ErrortStatisticProb. C2X12X22X32X42X52AR(1)Rsquared Mean dependent varAdjusted Rsquared . dependent var. of regression Akaike info criterionSum squared resid Schwarz criterionLog likelihood DurbinWatson statInverted AR Roots .34對模型做廣義一階差分回歸，得AR(1)=，經(jīng)差分后，得到新序列，設(shè)為 C3 Y3 X13 X23 X33 X43 X53生成新數(shù)據(jù)：C3=c2c2(1)*Y3=y2y2(1)*X13=x12x12(1)*X23=x22x22(1)*X33=x32x32(1)*X43=x42x42(1)*X53=x52x52(1)*補(bǔ)第一個數(shù)（1978年）：C3=c2*(^2)^Y3=y2*(^2)^X13=x12*(^2)^X23=x22*(^2)^X33=x32*(^2)^X43=x42*(^2)^X53=x52*(^2)^產(chǎn)生如下新序列：obsC3X13X23X33X43X53Y319781979198019811982198319841985198619871988198919901991199219931994199519961997199819992000
法律信息相關(guān)推薦
鄂ICP備17016276號-1