計(jì)量經(jīng)濟(jì)學(xué)(練習(xí)題)答案

11......第章簡(jiǎn)線回模2.1（）首先分析人均壽命與均GDP的量，用Eviews分：DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.CX1

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

有上可知關(guān)式為②關(guān)人均壽命與成人識(shí)字率的用Eviews分如專業(yè).專注

.22......DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.CX2

R-squared

MeanAdjustedS.E.regressionSum

S.D.Akaikecriterioncriterion

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

由上可知關(guān)式為③關(guān)人均壽命與一歲兒童疫苗種率的關(guān)用Eviews分如：DependentVariable:Y專業(yè).專注

.33......Method:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.CX3

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

由上可知關(guān)式為（）關(guān)人均壽命與人均模由可可系數(shù)為，說所建模型整體上對(duì)樣本數(shù)據(jù)擬合較。對(duì)于回歸系數(shù)的檢取查t布表得自由度為的臨界（紅色是自己加上的（

）對(duì)率系數(shù)的顯著性檢驗(yàn)表，1人均GDP對(duì)均壽命有顯著影。專業(yè).專注

.22......②關(guān)人均壽命與成人識(shí)字率模由上可，可決系數(shù)為說所建模型整體上對(duì)樣本數(shù)據(jù)擬合較對(duì)于回歸系數(shù)的t檢（

，對(duì)斜率系數(shù)的著性檢驗(yàn)表明成人識(shí)字率對(duì)人均壽命有顯著影③關(guān)人均壽命與一歲兒童疫苗模由可，可決系數(shù)為說所建模型整體上對(duì)樣本數(shù)據(jù)擬合較。對(duì)于回歸系數(shù)的檢：β3

對(duì)率系數(shù)的顯著性檢驗(yàn)表明一兒童疫苗接種率對(duì)人壽命有顯著影。2.2（）專業(yè).專注

.......①對(duì)浙江省預(yù)算收入與全省生總值的模用Eviews分結(jié)果如：DependentVariable:YMethod:SquaresDate:Time:(adjusted):1Includedobservations:adjustmentsVariableCoefficientStd.t-StatisticProb.X

C

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

②由可模型的參數(shù)斜率系數(shù)截為③關(guān)浙江省財(cái)政預(yù)算收入與全生產(chǎn)總值的模檢驗(yàn)?zāi)Ｐ偷娘@著：）可決系數(shù)為說所建模型整上對(duì)樣本數(shù)據(jù)擬合較專業(yè).專注

.......）對(duì)于回歸系數(shù)的檢：β對(duì)率系數(shù)的顯著性檢0.025驗(yàn)表明全省生產(chǎn)總值對(duì)財(cái)政預(yù)算總收入有顯著影。④用范形式寫出檢驗(yàn)結(jié)果如下—t=（）⑤經(jīng)意義是全生產(chǎn)總值每加1億，財(cái)政預(yù)算總收入增加億。（）當(dāng)x=32000時(shí)①進(jìn)點(diǎn)預(yù)測(cè)由可知代可：Y=②進(jìn)區(qū)間預(yù)：先由分：XMeanMedianMaximum專業(yè).專注

Y

.......Std.SkewnessKurtosis

Jarque-BeraProbability

SumSum

Observations由上表可知=∑X=2i

x

2

x(XX)—f

當(dāng)時(shí)將相關(guān)數(shù)據(jù)代入計(jì)算得Yf即的信區(qū)間為(3)對(duì)于浙江省預(yù)算收入對(duì)數(shù)與全省生產(chǎn)總值對(duì)數(shù)的模由分析結(jié)果如：DependentVariable:LNYMethod:Squares專業(yè).專注

.......Date:Time:(adjusted):1Includedobservations:adjustmentsVariableCoefficientStd.t-StatisticProb.

C

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

①模方程為lnY=0.980275lnX-1.918289②由可模型的參數(shù)斜率系數(shù)為截-③關(guān)浙江省財(cái)政預(yù)算收入與全生產(chǎn)總值的模檢驗(yàn)其顯著：）可決系數(shù)為說所建模型整上對(duì)樣本數(shù)據(jù)擬合較）對(duì)于回歸系數(shù)的t檢（β2

，對(duì)率系數(shù)的顯性檢驗(yàn)表明全省生產(chǎn)總值對(duì)財(cái)政預(yù)算總收入有顯著影。專業(yè).專注

.......④經(jīng)意全省生產(chǎn)總值每增長(zhǎng)財(cái)政預(yù)算總收入增長(zhǎng)專業(yè).專注

.......2.4（）對(duì)建筑面積建造單位成本模用分結(jié)果如下DependentVariable:YMethod:SquaresDate:Time:Sample:專業(yè).專注

.......Includedobservations:VariableCoefficientStd.t-StatisticProb.XC

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

由上可得建面積與建造成本回歸方程（）經(jīng)濟(jì)意建筑面積每增加1平方建筑單位成本每平方米減64.18400元（）①首進(jìn)行點(diǎn)預(yù)測(cè)由得當(dāng)，②再行區(qū)間估計(jì)用Eviews析專業(yè).專注

.......MeanMedianMaximumStd.SkewnessKurtosis

Y

XJarque-BeraProbability

SumSum

Observations由上表可知=∑X=2i

x

2

x(XX)f

當(dāng)，將相關(guān)數(shù)據(jù)代入計(jì)算得：Yf專業(yè).專注

.......即的信區(qū)間為3.1（）①對(duì)戶擁有家用汽車量計(jì)量經(jīng)模用分結(jié)果如專業(yè).專注

.......DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.X2

X3X4C

R-squared

MeanAdjustedS.E.regressionSum

S.D.Akaikecriterioncriterion

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

②得模型得23專業(yè).專注

.......③對(duì)型進(jìn)行檢驗(yàn)可決系數(shù)是修的可決系數(shù)為，說明模型對(duì)樣本擬合較好F檢驗(yàn)回歸方程顯。）檢驗(yàn)t統(tǒng)量分別為，均大于）所這些系數(shù)都是顯著的④依據(jù)可決系數(shù)越說擬合程度越F的與臨界值比若于臨界值則否定原假設(shè)回歸方是顯著若于臨界值則接受原假設(shè)回方程不。t的值與臨界值比較若于臨界值，否定原假設(shè)，系數(shù)都是顯著的若于臨界值則接受原假設(shè)系不顯著（）經(jīng)濟(jì)意人均Ｄ增１萬元百擁有家用汽車增加輛城鎮(zhèn)人口比重增加個(gè)分百戶擁有家用汽車減少輛交通工具消費(fèi)價(jià)格指數(shù)每上升１，百擁有家用汽車減少輛（）用析得DependentVariable:YMethod:SquaresDate:Time:Sample:專業(yè).專注

.......Includedobservations:VariableCoefficientStd.t-StatisticProb.X2

C

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

模型方程為-22.81005+1148.75823此分析得出的可決系數(shù)為擬程度得到了提可這樣改進(jìn)專業(yè).專注

.......3.2（１對(duì)口貨物總額計(jì)量經(jīng)濟(jì)用分析果如下：DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:專業(yè).專注

.......VariableCoefficientStd.t-StatisticProb.X2X3

C

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

①由可模型為Y+-2②對(duì)型進(jìn)行檢驗(yàn)可決系數(shù)是修的可決系數(shù)為說模型對(duì)樣本擬合較好F檢（回方程顯著t檢，統(tǒng)量分別為的系對(duì)應(yīng)值，大（，系數(shù)是顯著的X3的數(shù)對(duì)應(yīng)t為，小說此系數(shù)是不著。（）對(duì)于對(duì)數(shù)模用Eviews分結(jié)果如：專業(yè).專注

.......DependentVariable:LNYMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.

C

R-squaredAdjustedS.E.regressionSumLoglikelihood

MeanS.D.Akaikecriterion-1.296424SchwarzHannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

①由可模型為+1.7606952②對(duì)型進(jìn)行檢驗(yàn)）可決系數(shù)是修的可決系數(shù)說模型對(duì)樣本擬合較。專業(yè).專注

.......F檢）回歸方程顯t檢，統(tǒng)量分別-，均于t（）所以這些系數(shù)都是顯著（）①（式的經(jīng)濟(jì)意義工增加億元出貨物總額增加億人民幣匯率增加1，出口貨物總額增加18.85348億元②（式的經(jīng)濟(jì)意義工增加額每增出貨物總額增人幣匯率每增加，出貨物總額增1專業(yè).專注

.......3.3（）對(duì)家庭書刊費(fèi)對(duì)家庭月平均收入和戶主受教育年數(shù)計(jì)量模由分析結(jié)果如下DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.專業(yè).專注

.......XT

C

0.3279R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

①模為②對(duì)型進(jìn)行檢驗(yàn)可決系數(shù)是修的可決系數(shù)為說模型對(duì)樣本擬合較好F檢）回歸方程顯t檢統(tǒng)量分別為均大于（所以這些系數(shù)都是顯著的③經(jīng)意家庭月平均收入增加元家庭書刊年消費(fèi)支出增加，戶主受教育年數(shù)增加年家書刊年消費(fèi)支出增加元專業(yè).專注

.......（）用分析①DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.TC

0.8443R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)②DependentVariable:XMethod:Squares

Durbin-Watson

專業(yè).專注

.......Date:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.TC

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

以上分別是y與TX與T的一元回歸模型分別是YX（）對(duì)殘差進(jìn)行型分，用分析結(jié)果如：DependentVariable:專業(yè).專注

.......Method:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.E2C

3.96E-14

2.85E-15

R-squaredAdjustedS.E.regressionSum

MeanS.D.AkaikecriterionSchwarz

2.30E-14Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

模型為E+12參數(shù)斜系數(shù)為截距為3.96e-14（）由上可與的數(shù)是一樣?；貧w系數(shù)與被解釋變量的殘差系數(shù)是一樣，它們的變化規(guī)律是一致的專業(yè).專注

.......專業(yè).專注

.66......3.6（）預(yù)期的符號(hào),X,X12345

的符號(hào)為正的號(hào)為負(fù)（）根據(jù)Eviews分析到數(shù)據(jù)：DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.X2X3

X4

0.3346專業(yè).專注

.......X5X6

C

0.3984R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)①與期不相。

Durbin-Watson

②評(píng)價(jià)1可系數(shù)為數(shù)據(jù)相當(dāng)大可認(rèn)為擬合程很。2檢驗(yàn)）回歸方程顯3檢驗(yàn)X

,X1345，6

系數(shù)對(duì)應(yīng)的t值分別為，，均小于12）所以所得系數(shù)都是不顯著的（）根據(jù)Eviews分析到數(shù)據(jù)：DependentVariable:Y專業(yè).專注

.......Method:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.X5

2.20E-050.0000X6C

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

①得模型的方程-0.054965X56②評(píng)價(jià)1）可決系數(shù)為，數(shù)據(jù)相當(dāng)大可認(rèn)為擬合程度很2）F檢回歸方程顯著專業(yè).專注

.66......3）T檢驗(yàn)X5系對(duì)應(yīng)的值，大所系數(shù)是顯著的即人均對(duì)底存款余額有顯著影響X系數(shù)對(duì)應(yīng)的值-小于（，所以系數(shù)是不顯著的專業(yè).專注

.......4.3（）根據(jù)Eviews分析到數(shù)據(jù)：DependentVariable:LNYMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.LNGDP

C

R-squaredAdjusted

MeanS.D.專業(yè).專注

.......S.E.regressionSumLoglikelihood

Akaikecriterion-0.695670SchwarzHannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

得到的模型方程：LNGDP-3.111486tt（）①該型的可決系數(shù)為可系數(shù)很高F檢值為，明顯顯著。但當(dāng)時(shí)，），的數(shù)不顯著可能存在多重共線性②得相關(guān)系數(shù)矩陣如：LNGDP

LNGDP

，LNCPI間的相關(guān)系數(shù)很高證確實(shí)存在多重共線性（）由得專業(yè).專注

.......a）DependentVariable:LNYMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.LNGDP

C

R-squaredAdjustedS.E.regressionSumLoglikelihood

MeanS.D.Akaikecriterion-0.642056SchwarzHannan-Quinncriter.F-statisticProb(F-statistic)b)

Durbin-Watson

DependentVariable:LNYMethod:SquaresDate:Time:專業(yè).專注

.......Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.

C

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)c)

Durbin-Watson

DependentVariable:Method:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.專業(yè).專注

.......C0.0040R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

①得的回歸方程分別為L(zhǎng)NGDPttLNGDPt②對(duì)重共線性的認(rèn)識(shí)單方程擬合效果都很回系顯著判定系數(shù)較GDP對(duì)進(jìn)口的顯著的單一影響這兩個(gè)變量同時(shí)引入模型時(shí)影響方向發(fā)生了改變這只有通過相關(guān)系數(shù)的分析才能發(fā)現(xiàn)（）建議如僅僅是作預(yù)可不在意這種多重共線性但果是進(jìn)行結(jié)構(gòu)分還是應(yīng)該引起注意。專業(yè).專注

.......專業(yè).專注

.......4.4（）按照設(shè)計(jì)的論模，由分析：DependentVariable:CZSRMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.

GDPC

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

專業(yè).專注

.Loglikelihood

......Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

從回歸結(jié)果可見可系數(shù)為校的可決系數(shù)為模擬合的很好F的計(jì)量為，說在水平下回方程回歸方程整體上是顯著但t檢驗(yàn)結(jié)果表明國(guó)生產(chǎn)總值對(duì)財(cái)政收入的影顯，但回歸系數(shù)的符號(hào)為與實(shí)際不符合由可得知該程可能在多重共線。（）得到相關(guān)系矩陣如：GDP

GDP

由上表可知與，與GDP與之的相關(guān)系數(shù)都非常高說明確實(shí)存在多重共線性（）做輔助回歸被解釋變量

可決系數(shù)

方差擴(kuò)大因子專業(yè).專注

.......GDP

方差擴(kuò)大因子均大于存在嚴(yán)重多重共線性并通過以上分兩兩被解釋變量之間相關(guān)性都很高。（4解方式分作出財(cái)政收入與財(cái)政支出國(guó)生總值、稅總額之間的一元回歸專業(yè).專注

.......專業(yè).專注

.......5.2（）①用形法檢驗(yàn)繪制2的點(diǎn)用分如30,00025,00020,0002

15,00010,0005,00001,0001,5002,0002,500X

3,0003,5004,000由上圖可知模型可能存在異方，②Goldfeld-Quanadt檢）定義區(qū)間為1-7時(shí)由軟件分析：DependentVariable:YMethod:SquaresDate:Time:專業(yè).專注

.得e得e......Sample:Includedobservations:7VariableCoefficientStd.t-StatisticProb.TXC

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)1i

Durbin-Watson

）定義區(qū)間為時(shí)由件分析得DependentVariable:YMethod:SquaresDate:Time:專業(yè).專注

.得e得e......Sample:Includedobservations:7VariableCoefficientStd.t-StatisticProb.TX

C

0.9177R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)2i

Durbin-Watson

）根據(jù)Goldfeld-Quanadt檢統(tǒng)量：∑e22i

/∑e21i

在水平下分分母的自由度均為，查分布表得臨界值

（，因0.05，以接受原假，此檢驗(yàn)表明模型不存在異方差0.05（）存在異方差估計(jì)參數(shù)的方法專業(yè).專注

.......可以對(duì)模型進(jìn)行變換使用加權(quán)最小二乘法進(jìn)行計(jì)算得出模型方并對(duì)其進(jìn)行相關(guān)檢驗(yàn)③對(duì)型進(jìn)行對(duì)數(shù)變換進(jìn)分（）評(píng)價(jià)3.3所結(jié)論是可以相信，隨機(jī)擾動(dòng)項(xiàng)之間不存在異方回歸方程是顯著5.3(1)由Eviews件分析得DependentVariable:YMethod:SquaresDate:Time:Sample:專業(yè).專注

.......Includedobservations:VariableCoefficientStd.t-StatisticProb.XC

R-squared

MeanAdjustedS.E.regressionSum

S.D.Akaikecriterioncriterion

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

由上表可知年我國(guó)農(nóng)村居民家庭人均消費(fèi)支對(duì)均純收）的模型（）①由形法檢驗(yàn)專業(yè).專注

.......2E

6,000,0005,000,0004,000,0003,000,0002,000,0001,000,000002,000

6,000由上圖可知模型可能存在異方。②Goldfeld-Quanadt檢驗(yàn)）定義區(qū)間為1-12時(shí)由件分析得：DependentVariable:Method:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.X1C

0.6614R-squaredMean專業(yè).專注

.得e得e......AdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)1i

Durbin-Watson

）定義區(qū)間為時(shí)由件分析得DependentVariable:Method:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.X1C

R-squared

MeanAdjustedS.E.regression

S.D.Akaikecriterion

專業(yè).專注

.得e0.05得e0.05......SumLoglikelihood

SchwarzHannan-Quinncriter.F-statisticProb(F-statistic)2i

Durbin-Watson

）根據(jù)Goldfeld-Quanadt檢統(tǒng)量：∑e22i

/∑e21i

在水下分分母的自由度均為查布表得臨界值F

（，因0.05為

（，所以拒絕原假，此檢驗(yàn)表明模型存在異方。（））采用WLS估計(jì)過程，①用數(shù)建回：DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:WeightingW1VariableCoefficientStd.t-StatisticProb.X專業(yè).專注

.......C0.3389WeightedR-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-WatsonUnweighted

R-squared

MeanAdjustedS.E.regressionDurbin-Watson

S.D.Sumresid

對(duì)此模型進(jìn)行檢驗(yàn)得HeteroskedasticityF-statisticObs*R-squaredScaledSS

Prob.Prob.Prob.

專業(yè).專注

.......TestDependentVariable:Method:SquaresDate:Time:Sample:Includedobservations:CollinearregressorsspecificationVariableCoefficientStd.t-StatisticProb.C

WGT^2X*WGT^2

R-squaredMeanAdjustedS.D.dependentvar

S.E.regressionSum

AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

從上可知nR

比較計(jì)算的專業(yè).專注

統(tǒng)計(jì)量的臨界值因?yàn)閚R.......（）所接受原假，該模型消除了異方。估計(jì)結(jié)果為）（R2②用數(shù)2

用回歸分析得DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:WeightingW2VariableCoefficientStd.t-StatisticProb.X

C

R-squaredAdjusted

WeightedMeanS.D.專業(yè).專注

.......S.E.regressionSum

AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-WatsonUnweighted

R-squared

MeanAdjustedS.E.regressionDurbin-Watson

S.D.Sumresid

對(duì)此模型進(jìn)行檢驗(yàn)得HeteroskedasticityF-statisticObs*R-squaredScaledSS

Prob.Prob.Prob.

TestDependentVariable:Method:Squares專業(yè).專注

.......Date:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.C

WGT^2X^2*WGT^2

X*WGT^2

0.7816R-squaredMeanAdjustedS.D.dependentvar

S.E.regressionSum

AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

從上可知nR

比較計(jì)算的

統(tǒng)計(jì)量的臨界值因?yàn)閚R（）所接受原假，該模型消除了異方。估計(jì)結(jié)果為）（R2專業(yè).專注

.......③用數(shù)（x用歸分析：DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:WeightingW3VariableCoefficientStd.t-StatisticProb.X

C

R-squared

WeightedMeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-WatsonUnweighted

R-squaredMean專業(yè).專注

.......AdjustedS.E.regressionDurbin-Watson

S.D.Sumresid

對(duì)此模型進(jìn)行檢驗(yàn)得HeteroskedasticityF-statisticObs*R-squaredScaledSS

Prob.Prob.Prob.

TestDependentVariable:Method:SquaresDate:Time:Sample:Includedobservations:CollinearregressorsspecificationVariableCoefficientStd.t-StatisticProb.CWGT^2

專業(yè).專注

0.5869.......X^2*WGT^2R-squaredMeanAdjustedS.D.dependentvar

S.E.regressionSum

AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

從上可知nR

比較計(jì)算的

統(tǒng)計(jì)量的臨界值因?yàn)閚R

0.05（）所接受原假，該模型消除了異方。估計(jì)結(jié)果為）（R2經(jīng)過檢驗(yàn)發(fā)現(xiàn)用數(shù)的果最好所以綜上可即修改后的結(jié)果：）（R2專業(yè).專注

.......5.6(1)a)用Eviews模分析得DependentVariable:YMethod:SquaresDate:Time:專業(yè).專注

.......Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.XC

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

得回歸模型為：b)檢是否存在異方：①用Goldfeld-Quanadt檢如：）當(dāng)定義區(qū)間為1-13，由軟件分析：DependentVariable:YMethod:Squares專業(yè).專注

.得e得e......Date:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.X

C

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)1i

Durbin-Watson

）當(dāng)定義區(qū)間為1-13，由軟件分析：DependentVariable:YMethod:SquaresDate:Time:專業(yè).專注

.得e0.05得e0.05......Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.XC

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)2i

Durbin-Watson

）根據(jù)Goldfeld-Quanadt檢統(tǒng)量：∑e22i

/∑e21i

在水下分分母的自由度均為查布表得臨界值F

（，因0.05為F

（，所以拒絕原假，此檢驗(yàn)表明模型存在異方差②驗(yàn)用EViews軟分析得專業(yè).專注

.......HeteroskedasticityF-statisticObs*R-squaredScaledSS

Prob.Prob.Prob.

TestDependentVariable:RESID^2Method:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.C

X

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.專業(yè).專注

.......F-statisticProb(F-statistic)

Durbin-Watson

從上圖中可以看出nR

比計(jì)算的

統(tǒng)計(jì)量的臨界值因?yàn)閚R異方差

），所以拒絕原假設(shè)不絕備擇假設(shè)表明模型存在用以上兩種方法可檢驗(yàn)?zāi)Ｐ褪谴嬖诋惙讲頲)修模型）用加權(quán)二乘法修正異方差現(xiàn)象步驟如：①當(dāng)數(shù)時(shí)用軟件分析得DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:WeightingW1VariableCoefficientStd.t-StatisticProb.XC

專業(yè).專注

.R-squared

......WeightedMeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-WatsonUnweighted

R-squared

MeanAdjustedS.E.regressionDurbin-Watson得方程模型為：

S.D.Sumresid

）（R對(duì)此模型進(jìn)行檢驗(yàn)如：Heteroskedasticity專業(yè).專注

.......F-statisticObs*R-squaredScaledSS

Prob.Prob.Prob.

TestDependentVariable:Method:SquaresDate:Time:Sample:Includedobservations:CollinearregressorsspecificationVariableCoefficientStd.t-StatisticProb.C

WGT^2X*WGT^2

R-squared

MeanAdjustedS.E.regressionSum

S.D.Akaikecriterioncriterion

Loglikelihood

Hannan-Quinncriter.專業(yè).專注

.22......F-statisticProb(F-statistic)

Durbin-Watson

從上圖中可以看nR

比較計(jì)算的

統(tǒng)計(jì)量的臨界值因?yàn)閚R響

（）所接受原假設(shè)即模消除了異方差的影②當(dāng)數(shù)2

時(shí)用軟件分析得DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:WeightingW2VariableCoefficientStd.t-StatisticProb.XC

R-squared

WeightedMeanAdjustedS.E.regression

S.D.Akaikecriterion

專業(yè).專注

.......SumLoglikelihood

SchwarzHannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-WatsonUnweighted

R-squared

MeanAdjustedS.E.regressionDurbin-Watson得方程模型為：

S.D.Sumresid

）（R用檢模型：HeteroskedasticityF-statisticObs*R-squaredScaledSS

Prob.Prob.Prob.

專業(yè).專注

.......TestDependentVariable:Method:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.C

WGT^2X^2*WGT^2

X*WGT^2

0.2923R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

從上圖中可以看出nR

比計(jì)算的

統(tǒng)計(jì)量的臨界值因?yàn)閚R

），所以拒絕原假設(shè)不絕備擇假設(shè)表明模型存在專業(yè).專注

.......異方差此模型并未消除異方。③當(dāng)數(shù)時(shí)用件分析得DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:WeightingW3VariableCoefficientStd.t-StatisticProb.XC

R-squared

WeightedMeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

專業(yè).專注

.R-squared

......UnweightedMeanAdjustedS.E.regressionDurbin-Watson得方程模型為：

S.D.Sumresid

）（R對(duì)所得模型進(jìn)行檢驗(yàn)HeteroskedasticityF-statisticObs*R-squaredScaledSS

Prob.Prob.Prob.

TestDependentVariable:Method:SquaresDate:Time:專業(yè).專注

.......Sample:Includedobservations:CollinearregressorsspecificationVariableCoefficientStd.t-StatisticProb.C

WGT^2X^2*WGT^2

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

從上圖中可以看出nR

比計(jì)算的

統(tǒng)計(jì)量的臨界值因?yàn)閚R

），所以拒絕原假設(shè)不絕備擇假設(shè)表明模型存在異方差此模型并未消除異方。綜上所述用權(quán)二乘法w1效果最所模型：得方程模型為：專業(yè).專注

.......）（R）用對(duì)數(shù)模型法用軟件分析得：DependentVariable:LNYMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.C

R-squaredAdjustedS.E.regressionSumLoglikelihood

MeanS.D.Akaikecriterion-2.352753SchwarzHannan-Quinncriter.F-statisticProb(F-statistic)得到模型為

Durbin-Watson專業(yè).專注.

......對(duì)此模型進(jìn)行檢驗(yàn)得HeteroskedasticityF-statisticObs*R-squaredScaledSS

Prob.Prob.Prob.

TestDependentVariable:RESID^2Method:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.C

R-squaredAdjusted

MeanS.D.專業(yè).專注

.......S.E.regressionSumLoglikelihood

Akaikecriterion-7.192271SchwarzHannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

從上圖中可以看出nR

比計(jì)算的

統(tǒng)計(jì)量的臨界值因?yàn)閚R

）所接受原假設(shè)此型除了異方。綜合兩種方法改后的模型最：(2)）考慮價(jià)格因首先用軟件三者關(guān)系進(jìn)行分析如DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.X專業(yè).專注

.......PC

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

）用Goldfeld-Quanadt檢如下①當(dāng)本為1-13時(shí)進(jìn)回歸分析：DependentVariable:PMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.X專業(yè).專注

.得e得e......YC

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)1i

Durbin-Watson

②當(dāng)本為時(shí)做歸分析得DependentVariable:YMethod:SquaresDate:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.X

P

專業(yè).專注

.得e0.05得e0.05......CR-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)2i

Durbin-Watson

③根Goldfeld-Quanadt檢F統(tǒng)量∑e22i

/∑e21i

在水下分分母的自由度均為查布表得臨界值F

（，因0.05為F

（）所以拒絕原假此驗(yàn)表明模型存在異方。）用檢，軟件分析結(jié)果為HeteroskedasticityF-statisticObs*R-squaredScaledSS

Prob.Prob.Prob.

專業(yè).專注

.......TestDependentVariable:RESID^2Method:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.CX

X*PP

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson專業(yè).專注.

......從上圖中可以看出nR

比計(jì)算的

統(tǒng)計(jì)量的臨界值因?yàn)閚R在異方差

0.05（，以拒絕原假設(shè)不絕備擇假設(shè)表明模型存）修正①建對(duì)數(shù)模，用件分析如：DependentVariable:LNYMethod:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.C

R-squaredAdjustedS.E.regressionSumLoglikelihoodF-statistic

MeanS.D.Akaikecriterion-2.322188SchwarzHannan-Quinncriter.Durbin-Watson專業(yè).專注

.......Prob(F-statistic)對(duì)此模型進(jìn)行檢：HeteroskedasticityF-statisticObs*R-squaredScaledSS

Prob.Prob.Prob.

TestDependentVariable:RESID^2Method:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.C

專業(yè).專注

.

......R-squaredAdjustedS.E.regressionSumLoglikelihood

MeanS.D.Akaikecriterion-7.170690SchwarzHannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

從上圖中可以看出nR

比計(jì)算的

統(tǒng)計(jì)量的臨界值因?yàn)閚R

0.05（，以拒絕原假設(shè)不絕備擇假設(shè)表明模型存在異方差所此模型沒有消除方。②當(dāng)時(shí)用軟件分析如下DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:WeightingW1VariableCoefficientStd.t-StatisticProb.專業(yè).專注

.......XP

C

R-squared

WeightedMeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-WatsonUnweighted

R-squared

MeanAdjustedS.E.regressionDurbin-Watson

S.D.Sumresid

所得模型為對(duì)此模型進(jìn)W檢：專業(yè).專注

.......HeteroskedasticityF-statisticObs*R-squaredScaledSS

Prob.Prob.Prob.

TestDependentVariable:Method:SquaresDate:Time:Sample:Includedobservations:CollinearregressorsspecificationVariableCoefficientStd.t-StatisticProb.CWGT^2X*WGT^2

P^2*WGT^2P*WGT^2

R-squaredMean專業(yè).專注

.22......AdjustedS.E.regressionSum

S.D.Akaikecriterioncriterion

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

因?yàn)?/p>

（）所以接受原假設(shè)該型不存在異方差所以此模型消除了異方③當(dāng)用軟件分析得DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:WeightingW2VariableCoefficientStd.t-StatisticProb.XP

C

專業(yè).專注

.R-squared

......WeightedMeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-WatsonUnweighted

R-squared

MeanAdjustedS.E.regressionDurbin-Watson

S.D.Sumresid

所得模型為對(duì)該模型進(jìn)行檢驗(yàn)得HeteroskedasticityF-statisticObs*R-squared

Prob.Prob.專業(yè).專注.

......ScaledSSProb.TestDependentVariable:Method:SquaresDate:Time:Sample:Includedobservations:VariableCoefficientStd.t-StatisticProb.C

WGT^2

X^2*WGT^2X*WGT^2

P^2*WGT^2P*WGT^2

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

專業(yè).專注

.Loglikelihood

......Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

因?yàn)椋ǎ┧芙^原假設(shè)不絕備擇假，表明模型存在異方差所此模型沒有除異方。④當(dāng)時(shí)用軟件分析：DependentVariable:YMethod:SquaresDate:Time:Sample:Includedobservations:WeightingW3VariableCoefficientStd.t-StatisticProb.XP

C

0.5986R-squaredAdjusted

WeightedMeanS.D.專業(yè).專注

.......S.E.regressionSum

AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-WatsonUnweighted

R-squared

MeanAdjustedS.E.regressionDurbin-Watson

S.D.Sumresid

所得模型為對(duì)所得模型進(jìn)行檢得HeteroskedasticityF-statisticObs*R-squaredScaledSS

Prob.Prob.Prob.

Test專業(yè).專注

.......DependentVariable:Method:SquaresDate:Time:Sample:Includedobservations:CollinearregressorsspecificationVariableCoefficientStd.t-StatisticProb.CWGT^2

X^2*WGT^2P^2*WGT^2P*WGT^2

R-squared

MeanAdjustedS.E.regressionSum

S.D.AkaikecriterionSchwarz

Loglikelihood

Hannan-Quinncriter.F-statisticProb(F-statistic)

Durbin-Watson

因?yàn)椋ǎ┧芙^原假設(shè)不絕備擇假，表明模專業(yè).專注

.......型存在異方差所此模型沒有除異方。綜上所述修后的模型為Y=(-3.410502)R(3)體會(huì)對(duì)不同的模型可取對(duì)數(shù)模型或者加權(quán)二乘法對(duì)具有異方差性的模型進(jìn)行改進(jìn)而消除異方差但于不同的模，自度的不同，可能導(dǎo)致改進(jìn)的方法不同，所以要對(duì)改進(jìn)的模型進(jìn)行進(jìn)一步的檢驗(yàn)才專業(yè).專注

.......6.1(1)建居民收入消費(fèi)模，用分結(jié)果如下DependentVariable:YMethod:SquaresDate:Time:專業(yè).專注

.......Sample:In