陜西省建筑業(yè)總產(chǎn)值影響因素分析(編輯修改稿)
2025-07-21 14:17 本頁面
【文章內(nèi)容簡介】 4=*x2e53(1)Genr x3e54=*x3e53(1)Genr x4e54=*x4e53(1)Genr x5e54=*x5e53(1)補(bǔ)數(shù)據(jù)Ye54’1=*ye53ce54’1=*ce53x1e54’1=*x1e53x2e54’1=*x2e53x3e54’1=*x3e53x4e54’1=*x4e53x5e54’1=*x5e53產(chǎn)生新序列，如下：obsCE54YE54X1E54X2E54X3E54X4E54X5E541976197719781979198019811982198319841985198619871988198919901991199219931994199519961997199819992000200120022003200420052006200720082009對(duì)新序列進(jìn)行自相關(guān)及異方差檢驗(yàn)：Dependent Variable: YE54Method: Least SquaresDate: 06/10/11 Time: 15:41Sample: 1976 2009Included observations: 34VariableCoefficientStd. ErrortStatisticProb. CE54X1E54X2E54X3E54X4E54X5E54Rsquared Mean dependent varAdjusted Rsquared . dependent var. of regression Akaike info criterionSum squared resid Schwarz criterionLog likelihood DurbinWatson stat由上表可知，不存在自相關(guān)，下面進(jìn)行異方差檢驗(yàn)：檢驗(yàn)異方差：X1E54X2E54X3E54X4E54X5E54大小FstatisticF臨界值由檢驗(yàn)可知，即不存在自相關(guān)也不存在異方差，T,F檢驗(yàn)也合格，但x2e54與x3e54系數(shù)為負(fù)，即社會(huì)消費(fèi)品零售總額與外貿(mào)進(jìn)出口總值均與建筑業(yè)總產(chǎn)值呈負(fù)相關(guān)，這不符合經(jīng)濟(jì)意義，故舍去五元模型，尋求新的模型。六、 四元模型的自相關(guān)、異方差處理對(duì)Y C X1 X2 X4 X5進(jìn)行回歸Dependent Variable: YMethod: Least SquaresDate: 06/10/11 Time: 21:36Sample: 1997 2009Included observations: 13VariableCoefficientStd. ErrortStatisticProb. CX1X2X4X5Rsquared Mean dependent varAdjusted Rsquared . dependent var. of regression Akaike info criterionSum squared resid Schwarz criterionLog likelihood FstatisticDurbinWatson stat Prob(Fstatistic)查表可知dl= du= 則（ , ）范圍內(nèi)不存在自相關(guān)（0 ，）和（ ，4）范圍內(nèi)認(rèn)為存在自相關(guān)。，落在（ ，4）之中，所以存在自相關(guān)異方差檢驗(yàn)：利用夸特法對(duì)六元模型進(jìn)行異方差檢驗(yàn)將Xi按升序排列，i=1，2，4，5自由度為8，其檢驗(yàn)數(shù)據(jù)如下圖所示：x1x2x4x5大小FstatisticF臨界值如圖所示，所有的F值均大于臨界值，故存在異方差。因此，下面進(jìn)行消除異方差和自相關(guān)的處理第一次異方差處理：同樣采取與以上一樣的方法，各變量包括截矩項(xiàng)除以此時(shí)的殘差e，得到X1EX2EX4EX5EYE4，截距項(xiàng)的處理采用CE4=1/E4的方法Genr E4=abs(resid)Genr YE4=Y/E4Genr CE4=1/E4Genr X1E4=X1/E4Genr X2E4=X2/E4Genr X4E4=X4/E4Genr X5E4=X5/E4 產(chǎn)生的新序列數(shù)據(jù)如表21所示：obsCE4YE4X1E4X2E4X4E4X5E41976197719781979198019811982198319841985198619871988198919901991199219931994199519961997199819992000200120022003200420052006200720082009用新數(shù)據(jù)進(jìn)行回歸：Dependent Variable: YE4Method: Least SquaresDate: 06/11/11 Time: 09:34Sample: 1976 2009Included observations: 34VariableCoefficientStd. ErrortStatisticProb. CE4X1E4X2E4X4E4X5E4Rsquared Mean dependent varAdjusted Rsquared . dependent var. of regression Akaike info criterionSum squared resid Schwarz criterionLog likelihood DurbinWatson stat進(jìn)行異方差檢驗(yàn)：X1E4X2E4X4E4X5E4大小FstatisticF臨界值如圖可知，個(gè)F值均大于臨界值，仍然存在異方差，繼續(xù)進(jìn)行異方差處理：第二次異方差處理Genr E42=abs(resid)Genr YE42=YE4/E42Genr CE42=CE4/E42Genr X1E42=X1E4/E42Genr X2E42=X2E4/E42Genr X4E42=X4E4/E42Genr X5E42=X5E4/E42產(chǎn)生新數(shù)據(jù)如下：obsCE42YE42X1E42X2E42X4E42X5E421976197719781979198019811982198319841985198619871988198919901991199219931994199519961997199819992000200120022003200420052006200720082009對(duì)新數(shù)據(jù)進(jìn)行回歸：Dependent Variable: YE42Method: Least SquaresDate: 06/11/11 Time: 10:08Sample: 1976 2009Included observations: 34VariableCoefficientStd. ErrortStatisticProb. CE42X1E42X2E42X4E42X5E42Rsquared Mean dependent varAdjusted Rsquared . dependent var. of regression Akaike info criterionSum squared resid Schwarz criterionLog likelihood DurbinWatson stat進(jìn)行異方差檢驗(yàn)：X1E42X2E42X4E42X5E42大小FstatisticF臨界值如上表可知，依然存在異方差，繼續(xù)進(jìn)行異方差處理：第三次異方差處理：Genr E43=abs(resid)Genr YE43=YE42/E43Genr CE43=CE42/E43Genr X1E43=X1E42/E43Genr X2E43=X2E42/E43Genr X4E43=X4E42/E43Genr X5E43=X5E42/E43產(chǎn)生新數(shù)據(jù)：obsCE43YE43X1E43X2E43X4E43X5E4319761977197819791980198119821983198419851986198719881989199019911992199319941995
醫(yī)療健康相關(guān)推薦
鄂ICP備17016276號(hào)-1