試驗(yàn)三多元線性回歸模型的估計(jì)和檢驗(yàn)

1、廣東財(cái)建大琴舉商孕院HUASHANG COLLEGEGUANGDONG UN1VERS1TV OF FINANCES ECONOMICS實(shí)驗(yàn)報(bào)告課程名稱:計(jì)量經(jīng)濟(jì)學(xué)實(shí)驗(yàn)項(xiàng)目：實(shí)驗(yàn)三 多元線性回歸模型的估計(jì)和檢驗(yàn)實(shí)驗(yàn)類型：綜合性口設(shè)計(jì)性口 驗(yàn)證性專業(yè)班別： 姓 名: 學(xué) 號(hào): 實(shí)驗(yàn)課室:指導(dǎo)教師：石立實(shí)驗(yàn)日期：2014年5月12日廣東商學(xué)院華商學(xué)院教務(wù)處制、實(shí)驗(yàn)項(xiàng)目訓(xùn)練方案小組合作：是否小組成員：無(wú)實(shí)驗(yàn)?zāi)康模赫莆斩嘣€性回歸模型估計(jì)和檢驗(yàn)的方法。實(shí)驗(yàn)場(chǎng)地及儀器、設(shè)備和材料實(shí)驗(yàn)室：普通配置的計(jì)算機(jī)，Eviews軟件及常用辦公軟件實(shí)驗(yàn)訓(xùn)練內(nèi)容（包括實(shí)驗(yàn)原理和操作步驟）：【實(shí)驗(yàn)步驟】（一）國(guó)內(nèi)生產(chǎn)總值

2、的增長(zhǎng)模型：分析廣東省國(guó)內(nèi)生產(chǎn)總值的增長(zhǎng)，根據(jù)廣東數(shù)據(jù)（數(shù) 據(jù)見“表：廣東省宏觀經(jīng)濟(jì)數(shù)據(jù) -第三章.Xis ”文件，各變量的表示按照試驗(yàn)指導(dǎo)課本上的來(lái)表 示）選擇不變價(jià)GDP（ GDPB）、不變價(jià)資本存量（ZC）和從業(yè)人員（RY），把GDPB 作為因變量，ZC和RY作為兩個(gè)解釋變量進(jìn)行二元線性回歸分析。要求：按照試驗(yàn)指導(dǎo)課本Ro。Po2，分別作：1 作散點(diǎn)圖（GDPB同ZC，GDPB同RY）（結(jié)果控制在本頁(yè)）2進(jìn)行因果關(guān)系檢驗(yàn)（GDPB同ZC, GDPB同RY）（結(jié)果控制在本頁(yè)）Pairwise Granger Causality TestsDate: 05/12/14 Time: 12:10

3、Sample: 1978 2005Lags: 2Null Hypothesis:ObsF-StatisticProb.ZC does not Granger Cause GDPB263.849390.0376GDPB does not Granger Cause ZC19.07482.E-05Pairwise Granger Causality TestsDate: 05/12/14 Time: 12:10Sample: 1978 2005Lags: 3Null Hypothesis:ObsF-StatisticProb.RY does not Granger Cause GDPB252.88

4、7440.0641GDPB does not Granger Cause RY3.463090.0382從因果關(guān)系檢驗(yàn)看，ZC明顯影響GDPB, RY不明顯，這是可以理解的，計(jì)劃經(jīng)濟(jì) 時(shí)期存在著隱性失業(yè)，使得勞動(dòng)力的變化對(duì)產(chǎn)出的影響不明顯。3作GDPB同ZC和RY的多元線性回歸，寫出模型估計(jì)的結(jié)果，并分析模型檢驗(yàn)是均 否通過(guò)？（三個(gè)檢驗(yàn)）（結(jié)果控制在本頁(yè)）Dependent Variable: GDPBMethod： Least SquaresDate： 05/12/14 Time: 12:13Sample: 1978 2005Included obeervations: 23Coeffici

5、ent Std. Error t-StafiEtic ProbZC03771700.00835545.142650 0000RY0 3536890 0427573 2720280.0000C-600.5997113 7822-7 0362470 0000Rrsquared0 999152Mean dependent var1754112djustedR-squared0.999085S . dependent var1683.912S.E. of 佗gression50.94570.Akaike info critenon10.80035Sum squared resid648B6J1Schw

6、arz criterion10.94309Log likelihood-148.2050Hannan-Quinn enter.10.S4399F-staiistlc1473632Durbin-Wais on stat0 443992ProbfF-stalistic0.000000得到的估計(jì)方程GDPB=0.377170*ZC+0.353689*RY-800.59974.將建立的二元回歸模型 （GDPB同ZC和RY）同一元回歸模型（GDPB同ZC、GDPB同RY）相比較，分析優(yōu)點(diǎn)。（結(jié)果控制在本頁(yè)）Dependent Variable: GDPB Method: Least Squares Oa

7、te:05；12/14 Time: 12:16 Sample 197B 2005 included observations: 28Co efficientStd Errort-StatlsticProb.ZC C0.44289813397210004S96904600025.570545.2393140.00000.0000R-squared Adjusted R-s qua red S.E of regression Sum squared resid Log likelihood F-statistic Prob（F-statistic）0.9968330.99671196 573022

8、4248&.Q -166.6611S183.0110000000Mean dependent var S.D.dependent var Akaike info criterion Sctiwarz aiteilon Hannan-Quinn criter. Durtoi n-Watson stat1764 112168391212.0472212 1423912.076320 167556Dependent Variable: GDPB Meihod: Least Squares Date:05/12/14 Time: 12:26Sample. 197S 2005Included o

9、bservations:CoefficientSid Errort-StatistiGProt>,RY2.1893170.11773518.595230 0000C-5519.127400 4253-13.793190 0000R-sq uared0.9SOT67Iean dependentvsr1754112Adjusted F?-squared0.927377£.D. dependent var1633 912S.E. olrectression4537907Akai Ke info crilerion15.14190Sum squared rasid5354077Schw

10、arz criterion1523706Log likelihood-209 9866Hannan-Quinn enter.15.17099F-stathtc345 7S44Durtin-Watson stat0.07S643Proh(F-statistic)0.000000（結(jié)果控制在本頁(yè)） 方程顯著性F檢驗(yàn)顯著。5. 結(jié)合相關(guān)的經(jīng)濟(jì)理論，分析估計(jì)的二元回歸模型的經(jīng)濟(jì)意義 估計(jì)方程的判定系數(shù)人R2接近1;參數(shù)顯著性t檢驗(yàn)均大于2； 調(diào)整的判定系數(shù)為0.99085,比上面的一元回歸有明顯改善。（二）宏觀經(jīng)濟(jì)模型：根據(jù)廣東數(shù)據(jù)，研究廣東省居民消費(fèi)行為、固定資產(chǎn)投資行為、 貨物和服務(wù)凈出口行為和存

11、貨行為，分別建立居民消費(fèi)模型、固定資產(chǎn)投資模型、貨 物和服務(wù)凈出口模型和存貨增加模型。要求：按照試驗(yàn)指導(dǎo)課本Ro5Rl2，分別作出以下模型，并對(duì)需要改進(jìn)的模型 進(jìn)行改進(jìn)。寫出最終估計(jì)的模型結(jié)果，并結(jié)合相關(guān)的經(jīng)濟(jì)理論，分析模型的經(jīng)濟(jì)意義。（數(shù)據(jù)見“表：廣東省宏觀經(jīng)濟(jì)數(shù)據(jù) -第三章.XlS ”文件，各變量的表示按照試驗(yàn)指導(dǎo)課本上的 來(lái)表示。）1 居民消費(fèi)模型（結(jié)果控制在本頁(yè)） 根據(jù)經(jīng)濟(jì)理論居民消費(fèi)XFJ取決于勞動(dòng)報(bào)酬LB，看散點(diǎn)圖和因果關(guān)系檢驗(yàn)。10,00098,000*6,000 _*醫(yī)4,000 _*K2,000 -*«*0 -02,0004,0006,0008,00010,000L

12、BPairwise Granger Causality TestsDate: 05/12/14 Time: 12:34Sample: 1978 2005Lags: 2Null Hypothesis:ObsF-StatisticProb.LB does not Granger Cause XFJ267.190100.0042XFJ does not Granger Cause LB5.455160.0124從散點(diǎn)圖看它們之間具有線性關(guān)系，從因果關(guān)系檢驗(yàn)看它們之間似乎具有雙向因果關(guān) 系。宏觀經(jīng)濟(jì)中確實(shí)如此。進(jìn)行一元線性回歸如下：Dependent Variable: XFJMethod: Leas

13、t SquaresDate: 06/12/14 Time: 12:36Sample: 1973 2005Induced observati ons: 2SCaefficientStd Errort-StatisticProb.L60 986702001691659.330100.0000c-75.9966259.99073-1 2658060 2165R-squared0.992416LTean dependent var236 2 277Adjusted R-squared0 992125S D. dependent var2565.722S.e. or regression227.6&am

14、p;09Akai Ke info criterion13.76260Sum squared resid1347921.Schwarz criterion13 85775Log likelihood-190.6765Hannan'Quinn enter13.79169F'Statistic2402401Durbin-V-atson slat0.701578Pro bF-stati stic)0 000000得到回歸方程XFJ=0.986702*LB-75.996622 固定資產(chǎn)投資模型（結(jié)果控制在本頁(yè)）固定資產(chǎn)投資TZC顯然取決于固定資產(chǎn)折舊ZJ、營(yíng)業(yè)盈余丫丫和財(cái)政支出CZ,進(jìn)行

15、三 元線性回歸如下：Dependent Variable: T2GMethod: Least SquaresDate: 05/12/14 Time: 12:45Sample： 1973 2005Included observations: 28CoeflidentStd Errort-Stati sticProb.ZJ11118640.2431524.5727160.0001YY0 4316920.0525663 2123520.0000CZ01432100 4053QB0.3533380 7269C31.2762527.825171 1240270.2721R-squared0.997573

16、Me a n depe1628997Adjusted R-squared0.997270S.D. dependent var2003852S.E. of regression104.7010Akai Ice info criterion1227166Sum squared resid2630951Schwarz criterion1246197Log likelihood-167.8032Hannan-Quinn criter1232984F-statistic3288.646Curb in-Wats on stat1 298515Pro b(F-stati stic)0.000000分別去掉

17、一個(gè)解釋變量進(jìn)行三個(gè)二元線性回歸如下：Dependent Variable:TZGMethod: Least SquaresDate: 05/12/14 Time: 12:47Sample: 1978 2005Induded observations: 2SCoeffidentStd. Error(-StatisticProb.ZJ1.1918780.08699313 700910 0000YY0.4384220.D481299.1093&50,0000C33.6561326.520921.2590410.2161R-squared0.997561Mean dependentvar16

18、29.997Adjusted R-squared0.997366S.D. dependent war2003 952S.E of regression102 S521Akaike info criterion12.20542Sum squared resid2644637Sdiwarz criterion12.34B15Log likelihood-167.8758Hannan-Quinn alter.12.24905F-statistic5111.852Durbin-'/Jatson stat1 370345ProtHF-statistic)0000000Dependent vari

19、able: tzgHettiod: Least SquaresDate: 05712/14 Time: 12:48Sample: 1978 2005Included observations: 29Co efficientStd ErrorbStati sticProb.ZJ1.09857S0.4650212.3624280.0262CZ1.3493010.7224791 3676010.0736C-45.6139450.11293-0 91022303714R*squared0 990754Mean dependentvar1628997Adjusted R-squared0990014S.

20、D. dependent var2003852S E. of regression200.2421Akaike info criterio n13.53739Sum squared resid1002422Schwarz criterion136B062Log likelihood-186.5304Hannan-Quinn criter.13 58152F-statistic1339.431Durbin-Watson statQ.436795Prob (F-statistic)0.000000Dependent Variable. TZG Method: Least Squar&sDa

21、te: 05/12/14 Time: 12:49Sample' 1978 2005Included observations: 28CoefficientSid. Error1-SlatjsticPros.YY CZ0,4300930 070453S1C4709000001.3&92760 197S469 W31350.0000G2C 9169337 170150 5627360 5736R*squared0.995459Mean aependent var1626997Adjusted R-sciuareti0.9S5096SO dependentvar2003852S.E

22、cfregression140.5301Aka ike info Gfiierion12A2633Sum squared resid452313.8Schwarz cr iterign12.96957Log likelihood-17S.f756Hannan-Quinn crikr,12.S7Q47F-statistic2740.226Durbin AVatson stat0.761924Prostic；0.000000從上面三個(gè)回歸結(jié)果可以看出，只要固定資產(chǎn)折舊ZJ和財(cái)政支出CZ其中一個(gè)不在方程中，回歸就能得到很好的擬合?，F(xiàn)在暫且去最后一個(gè)回歸方程來(lái)使用，方程為TZG=0.430093*YY

23、+1.869278*CZ+20.918933 貨物和服務(wù)凈流出模型（結(jié)果控制在本頁(yè)）先考慮影響貨物和服務(wù)凈流出CK的因素為支出法的國(guó)內(nèi)生產(chǎn)總值 GDP，看散點(diǎn)圖和 因果關(guān)系檢驗(yàn)。Pairwise Granger Causality Tests Date: 05/12/14 Time： 13:00 Sample: 197S 2005Lags: 2Null Hypothesis:ObsF-StatisticProbCK does not Granger Cause GDP GDP does not Granger Cause CK2610.25637 44017O.OOQ8Q.0036從散點(diǎn)圖和因

24、果關(guān)系檢驗(yàn)看它們具有關(guān)系，進(jìn)行一元線性回歸如下:Depende nt Variable: GDP Method: L&ast Squares Date: 05/12/14 Time: 13:02Sample: 1979 2005 Included oBsenrations:28Co efficientStd. Errort-StatisticProbCK9 3122230.5205S117.88813O.OQ9QC1450.714400,22653.&497160.0012R-5quared0.924852M&an depend&rit'ar5437,3

25、86Adjusted R-squareti0 921962S D depen dent var&3M032S E or regression1761.049Akaike info criterion17.35396Sum squared residi80633654Schwarz criterion17.94911Log liKAlitiaod*247 9554HannanQuinn critei.17 30395F-statistic319.9853Duroin-Z/atson stat0.868719Pro b(F-statstc)OOOOOQQ在所有收集到的統(tǒng)計(jì)數(shù)據(jù)中，年利率LL

26、是一個(gè)可以考慮引入的因素，引入LL進(jìn)行二元線性回歸如下：Dependentvariable: CKMethod： Least SquaresDate: 05/12/14 Time: 13:04Sample: 1978 2005Included observations: 28CoefficientStd. Errort-StatisticProb.GDP0.088239Q 00552515.970570 0000LL-42.6598911.83064-3 6058800.0014C202.2173952503821230080 0438R-squared0.950564Mean depende

27、ntvar427 0379Adjusted R-squared0.945&09S.D. dependentvar651.0303S.E of regression150.4304Akai Ke info criterion12.96584Sum squared resid565732.7Schwarz criterion13 10857Log likelihood-179.5217Hannan-Quinn crrter.13.00947F-statistic2403512Durbin-Watson stat1,504205Prob (F-stati Stic)0.000000最后得到回

28、歸方程CK=0.88239*GDP-42.65989*LL+202.21734 存貨增加模型（結(jié)果控制在本頁(yè)）存貨增加TZC顯然取決于城鄉(xiāng)儲(chǔ)蓄CX和商品零售價(jià)格指數(shù)PSL,進(jìn)行二元線性回歸如下：Dependent Variable: TZC Method: Least Squares D3te:C5/12/i4 Time: 13:12 Sample' 1978 2005 Included observations: 2ECoefficientStJ. Errort-StatisticPro&.CX0.0306330.004736 46369S0 0000PSL178080601

29、906593 9551120 0000G-2D9 054645E4519-4 5600130.0001R-squared0.952473Mean dependemt var4243629Adjusted R-squared0.946671S.D. dependent var3922360S.E. of regressionSS.S6445Akaike info criterion11,91305Sum squared resid!197422.3Schwarz criterion12 05579Log likelihood-1637826Hannan-Quinn enter.11.95669F-Btatistic250.5102Durbin-Watsnn stat2164713Pro b(F-statist