（3）可決系數(shù)R2=0.958316表示在財(cái)政收入Y的總變差中由模型作出的解釋部分ESS=RSS*R2/(1-R2)=(1.91E+08)*0.958316/(1-0.958316F=(n-2)ESS/RSS,ESS=F*RSS/(n-2)=4.DependentVariable:YMethod:LeastSquaresDate:04/10/10Time:17:31Sample:19782007Includedobservations:30VariableCoefficientt-StatisticC-1354.856655.7254-2.0661940.0482X0.1796720.00708225.371520.0000R-squared0.958316Meandependentvar10049.04AdjustedR-squared0.956827S.D.dependentvar12585.51S.E.ofregression2615.036Akaikeinfocriterion18.64028Sumsquaredresid1.91E+08Schwarzcriterion18.73370Loglikelihood-277.6043F-statistic643.7141Durbin-Watsonstat0.235088Prob(F-statistic)0.000000DependentVariable:XMethod:LeastSquaresDate:04/19/10Time:22:03Sample:19782008Includedobservations:31VariableCoefficientt-StatisticT7085.937571.777312.392830.0000C-46361.2510480.85-4.4234240.0001R-squared0.841167Meandependentvar67013.75AdjustedR-squared0.835690S.D.dependentvar70245.95S.E.ofregression28474.28Akaikeinfocriterion23.41373Sumsquaredresid2.35E+10Schwarzcriterion23.50625Loglikelihood-360.9128F-statistic153.5822令t=2008，其預(yù)測(cè)結(jié)果X=173302.807747再根據(jù)X對(duì)Y進(jìn)行預(yù)測(cè)，其預(yù)測(cè)結(jié)果為Y=29782.7237932X2008=173302.683609Y2008=29782.7707474[Y0-tα/2Sê(e0),Y0+tα/2Sê(e0)]=[29782.P124,6.該家庭在衣著用品方面的開DependentVariable:YMethod:LeastSquaresDate:04/20/10Time:09:24Sample:19912000Includedobservations:10VariableCoefficientStdErrort-StatisticProbC-3.7554552.679575-1.4015110.2038X10.1838660.0289736.3460710.0004X20.3017460.1676441.7999230.1149R-squaredAdjustedR-squared0.9606160.949364MeandependentvarS.D.dependentvar8.0800003.724931S.E.ofregression0.838204Akaikeinfocriterion2.728214Sumsquaredresid4.918099Schwarzcriterion2.818990Loglikelihood-10.64107F-statistic85.36888Durbin-Watsonstat2.725104Prob(F-statistic)0.000012EQ\*jc3\*hps36\o\al(\s\up5(^),Y)X2iEQ\*jc3\*hps36\o\al(\s\up6(^),β)EQ\*jc3\*hps36\o\al(\s\up6(^),β)EQ\*jc3\*hps36\o\al(\s\up7(^),β)影響顯著。從經(jīng)濟(jì)意義上分析，衣著用品作為日?；鞠M(fèi)品，其開支必然會(huì)與總開支保持一定比例的同步增長(zhǎng)。22無法否定原假設(shè)H0，EQ\*jc3\*hps36\o\al(\s\up7(^),β)2在統(tǒng)計(jì)上不顯著，即就單獨(dú)而言，衣著用品價(jià)格（X2）對(duì)衣著用品方面的開支（Y）影響不顯著。從經(jīng)濟(jì)意義上分析，衣著用品的需求量具有一定彈性，消費(fèi)者在衣著用品方面的開支主要由收入決定，當(dāng)商品價(jià)格發(fā)生變化時(shí)消費(fèi)者會(huì)調(diào)節(jié)需求量使衣著用品方面的開支在總開支中保持一定比例，因此當(dāng)總開支不變時(shí)，衣著用品價(jià)格（X2）變動(dòng)對(duì)衣著用品方面的開支（Y）影響不顯著。EQ\*jc3\*hps36\o\al(\s\up6(^),β)EQ\*jc3\*hps36\o\al(\s\up6(^),β)EQ\*jc3\*hps36\o\al(\s\up7(^),β)受備擇假設(shè)H1，即模型的整體擬合優(yōu)度較好，總開支（X1）和衣著用品價(jià)格（X2）對(duì)衣著用品方面的開支（Y）的共同影響顯著。7.(1)倒數(shù)回歸模型DependentVariable:YMethod:LeastSquaresDate:05/26/10Time:17:22Includedobservations:12VariableCoefficientStdErrort-StatisticC-0.2594371.008640-0.2572140.8022X120.587884.6794824.3996070.0013R-squaredAdjustedR-squaredS.E.ofregressionSumsquaredresidLoglikelihoodDurbin-Watsonstat0.6593600.6252960.7783866.058842-12.926940.639368MeandependentvarS.D.dependentvarAkaikeinfocriterionSchwarzcriterionF-statisticProb(F-statistic)4.0666671.2716012.4878232.56864019.356540.001336(2)線性回歸模型DependentVariable:YMethod:LeastSquaresDate:05/11/10Time:21:29Includedobservations:12VariableCoefficientStd.Errort-StatisticC8.0147011.2401886.4624920.0001X-0.7882930.241772-3.2604790.0086R-squared0.515286Meandependentvar4.066667AdjustedR-squared0.466815S.D.dependentvar1.271601S.E.ofregression0.928517Akaikeinfocriterion2.840556Sumsquaredresid8.621445Schwarzcriterion2.921374Loglikelihood-15.04334F-statistic10.63073Durbin-Watsonstat0.657106Prob(F-statistic)0.008567DependentVariable:Y1Method:LeastSquaresDate:05/26/10Time:22:06Sample:19912007Includedobservations:17VariableCoefficientStdErrort-StatisticProbC-7.7352122.462366-3.1413740.0078T10.0141800.0105631.3424400.20241.0821320.3853772.8079840.01480.9539830.05582017.090260.0000R-squared0.998773Meandependentvar5.600728AdjustedR-squared0.998490S.D.dependentvar0.749821S.E.ofregression0.029133Akaikeinfocriterion-4.031540Sumsquaredresid0.011034Schwarzcriterion-3.835490Loglikelihood38.26809F-statistic3528.578Durbin-Watsonstat1.557960Prob(F-statistic)0.000000DependentVariable:YMethod:LeastSquaresDate:05/26/10Time:22:35Sample:19912007Includedobservations:17Convergencenotachievedafter100iterationsY=C(1)*(1+C(2))^T*L^C(3)*K^C(4)Coefficientt-StatisticC(1)0.8104910.7832701.0347540.3196C(2)0.5490190.01575134.855540.0000C(3)0.6534700.1780033.6711100.0028C(4)0.8407430.05086816.527890.0000R-squared0.999991Meandependentvar9687017.AdjustedR-squared0.999989S.D.dependentvar20134768S.E.ofregression67362.50Akaikeinfocriterion25.27589SumsquaredresidLoglikelihood5.90E+10-210.8451SchwarzcriterionDurbin-Watsonstat25.471941.794933DependentVariable:LOG(Y)Method:LeastSquaresDate:05/26/10Time:23:00Sample:19912007Includedobservations:17Convergencenotachievedafter100iterationsLOG(Y)=LOG(C(1))+T*LOG(1+C(2))+C(3)*C(4)*LOG(L)+C(3)*(1-C(4))*LOG(K)-(1/2)*C(3)*C(4)*C(5)*(1-C(4))*(LOG(K/L))^2Coefficientt-StatisticC(1)3.1241561.0308913.0305390.0105C(2)0.5100550.003428148.79100.0000C(3)1.3100600.05483123.892650.0000C(4)0.2826570.02296812.306430.0000C(5)-0.1039630.055655-1.8680010.0864R-squaredAdjustedR-squaredS.E.ofregressionSumsquaredresidLoglikelihood0.9999990.9999990.0032090.00012476.44661MeandependentvarS.D.dependentvarAkaikeinfocriterionSchwarzcriterionDurbin-Watsonstat13.599202.702497-8.405483-8.1604210.943773），it8.008442-0.066759）R2=0.000186df=26F=0.004457(P=0.947327)DW=0.876171ARCHTest:F-statisticObs*R-squared37.7408918.77622ProbabilityProbability0.0000000.000084i因此可以取權(quán)重變量W1=Y0.015123。EQ\*jc3\*hps36\o\al(\s\up5(^),S)EQ\*jc3\*hps30\o\al(\s\up1(^),E)EQ\*jc3\*hps36\o\al(\s\up5(^),S)EQ\*jc3\*hps30\o\al(\s\up1(^),E)EQ\*jc3\*hps36\o\al(\s\up9(^),S)EQ\*jc3\*hps30\o\al(\s\up1(^),E)EQ\*jc3\*hps36\o\al(\s\up8(^),S)EQ\*jc3\*hps30\o\al(\s\up1(^),E)④擬合優(yōu)度均有所提高，而且懷特檢驗(yàn)均認(rèn)為消較，可以看出模型④在擬合優(yōu)度、p值及相關(guān)系數(shù)方面均優(yōu)于模型③,因此選擇LUL隨機(jī)擾動(dòng)項(xiàng)存在正自相關(guān)性。EQ\*jc3\*hps35\o\al(\s\up2(^),P)EQ\*jc3\*hps35\o\al(\s\up2(^),P)NY=-103.5852+1.482130NXSE37.909010.218745）t-2.7324706.775618）P0.01250.0000）R2=0.686141DW=1.583575F=45.90900P（F0.000001）所以通過廣義差分法對(duì)模型消除一階自相關(guān)性后的輸出結(jié)果如下：EQ\*jc3\*hps36\o\al(\s\up5(^),Y)(2)求各解釋變量間的簡(jiǎn)單相關(guān)系數(shù)結(jié)果如下表所示：從該矩陣中可初步判斷，各解釋變量之間存在多重共線性?？梢钥闯鯴1、X4、與Y高度相關(guān)，即農(nóng)業(yè)化肥施用量、農(nóng)業(yè)機(jī)械動(dòng)力、與糧食總產(chǎn)量高度相關(guān)；X5與Y較為相關(guān)，即農(nóng)業(yè)勞動(dòng)力與糧食總產(chǎn)量較為相關(guān)。下面利用輔助回歸方程的可決系數(shù)R2和F統(tǒng)計(jì)量檢驗(yàn)解釋變量間的多重共線性:;;;;;VIFiX1、X4的VIFi值大于10，說明存在較嚴(yán)重的多重共線性根據(jù)以上分析表明,各個(gè)解釋變量Xi與其他解釋變量的線性組合在整體上具有較高的相關(guān)程度,即各解釋變量間存在多重共線性。2所以最終引入變量X1,X3,X5，采用逐步回歸法消除解釋變量間的多重共線性，最終得EQ\*jc3\*hps36\o\al(\s\up5(^),Y)EQ\*jc3\*hps36\o\al(\s\up5(^),Y)EQ\*jc3\*hps36\o\al(\s\up5(^),δ)EQ\*jc3\*hps36\o\al(\s\up5(^),β)*2EQ\*jc3\*hps36\o\al(\s\up5(^),Y)t其經(jīng)濟(jì)意義為：全地區(qū)工業(yè)增加值（X）每增加一個(gè)單位，期望的國(guó)有企業(yè)基本建設(shè)新首先采用工具變量法將Y對(duì)X的滯后值進(jìn)行回歸，再將滯后值Y其代入原模型，并消除一將所得結(jié)果帶入適應(yīng)性期望模型其經(jīng)濟(jì)意義為：當(dāng)年預(yù)期的全地區(qū)工業(yè)增加值X每增加一個(gè)單位將導(dǎo)致當(dāng)年該地區(qū)國(guó)有企業(yè)基本建設(shè)新增固定資產(chǎn)Y增加0.012883個(gè)單位。D1i10第一季度其他季度D2i10第二季度D3i其他季度（1）當(dāng)認(rèn)為季度影響是利潤(rùn)平均值發(fā)生變異，因以加法的方式引入虛擬變量，即：在該情況下估計(jì)的利潤(rùn)模型為：EQ\*jc3\*hps36\o\al(\s\up5(^),Y)SE:(1892.072)(0.011481)(654.2925)(630.7096)(636.1179)t:(3.632113)(3.331281)(-0.281000)(1.805959)(-0.631427)P:(0.0018)(0.0035)(0.7817)(0.0868)(0.5353)2=0.525494（2）當(dāng)認(rèn)為季度影響利潤(rùn)關(guān)于銷售額的變化率發(fā)生變異，因以乘法的方式引入虛擬變量：0在該情況下估計(jì)的利潤(rùn)模型為：EQ\*jc3\*hps36\o\al(\s\up5(^),Y)SE:(1753.300)(0.011139)(0.004259)