EViews計(jì)量經(jīng)濟(jì)學(xué)實(shí)驗(yàn)報(bào)告-多重共線性的診斷與修正的討論(共11頁(yè))

1、精選優(yōu)質(zhì)文檔-傾情為你奉上實(shí)驗(yàn)題目 多重共線性的診斷與修正 一、實(shí)驗(yàn)?zāi)康呐c要求：要求目的：1、對(duì)多元線性回歸模型的多重共線性的診斷； 2、對(duì)多元線性回歸模型的多重共線性的修正。二、實(shí)驗(yàn)內(nèi)容根據(jù)書上第四章引子“農(nóng)業(yè)的發(fā)展反而會(huì)減少財(cái)政收入”，19782007年的財(cái)政收入，農(nóng)業(yè)增加值，工業(yè)增加值，建筑業(yè)增加值等數(shù)據(jù)，運(yùn)用EV軟件，做回歸分析，判斷是否存在多重共線性，以及修正。三、實(shí)驗(yàn)過(guò)程：（實(shí)踐過(guò)程、實(shí)踐所有參數(shù)與指標(biāo)、理論依據(jù)說(shuō)明等）（一）模型設(shè)定及其估計(jì)經(jīng)分析，影響財(cái)政收入的主要因素，除了農(nóng)業(yè)增加值，工業(yè)增加值，建筑業(yè)增加值以外，還可能與總?cè)丝诘纫蛩赜嘘P(guān)。研究“農(nóng)業(yè)的發(fā)展反而會(huì)減少財(cái)政收入”這

2、個(gè)問(wèn)題。設(shè)定如下形式的計(jì)量經(jīng)濟(jì)模型：=+其中，為財(cái)政收入CS/億元；為農(nóng)業(yè)增加值NZ/億元；為工業(yè)增加值GZ/億元；為建筑業(yè)增加值JZZ/億元；為總?cè)丝赥POP/萬(wàn)人；為最終消費(fèi)CUM/億元；為受災(zāi)面積SZM/千公頃。圖1： 19782007年財(cái)政收入及其影響因素?cái)?shù)據(jù)年份財(cái)政收入CS/億元農(nóng)業(yè)增加值NZ/億元工業(yè)增加值GZ/億元建筑業(yè)增加值JZZ/億元總?cè)丝赥POP/萬(wàn)人最終消費(fèi)CUM/億元受災(zāi)面積SZM/千公頃19781132.31027.51607138.2962592239.15079019791146.41270.21769.7143.8975422633.73937019801159

3、.91371.61996.5195.5987053007.94452619811175.81559.52048.4207.13361.53979019821212.31777.42162.3220.73714.833130198313671978.42375.6270.64126.43471019841642.92316.12789316.74846.33189019852004.82564.43448.7417.95986.344365198621222788.73967525.76821.84714019872199.432334585.8665.87804.64209019882357.

4、23865.45777.28109839.55087019892664.94265.9648479411164.24699119902937.150626858859.412090.53847419913149.485342.28087.11015.114091.95547219923483.375866.610284.5141517203.35133319934348.956963.8141882266.521899.94882919945218.19572.719480.72964.729242.25504319956242.212135.824950.63728.836748.24582

5、119967407.9914015.429447.64387.443919.54698919978651.1414441.932921.44621.648140.65342919989875.9514817.634018.44985.851588.250145199911444.081477035861.55172.155636.949981200013395.2314944.7400365522.36151654688200116386.0415781.343580.65931.766878.352215200218903.641653747431.36465.571691.24711920

6、0321715.2517381.754945.57490.877449.554506200426396.4721412.7652108694.387032.937106200531649.292242076912.910133.896918.138818200638760.22404091310.911851.1.341091200751321.7828095.214014.1.648992利用EV軟件，生成、等數(shù)據(jù)，采用這些數(shù)據(jù)對(duì)模型進(jìn)行OLS回歸。（二）診斷多重共線性1、雙擊“Eviews”，進(jìn)入主頁(yè)。輸入數(shù)據(jù)：點(diǎn)擊主菜單中的File/Open /EV WorkfileExcel多重共線性

7、的數(shù)據(jù).xls ；2、在EV主頁(yè)界面的窗口，輸入“l(fā)s y c x2 x3 x4 x5 x6 x7”，按“Enter”.出現(xiàn)OLS回歸結(jié)果,圖2： 圖2： OLS 回歸結(jié)果Dependent Variable: YMethod: Least SquaresDate: 10/12/10 Time: 17:07Sample: 1978 2007Included observations: 30VariableCoefficientStd. Errort-StatisticProb.  C-6646.6946454.156-1.0.3138X2-0.0.-2.0.0074X31.

8、0.4.0.0001X4-2.2.-1.0.1963X50.0.1.0.2653X6-0.0.-0.0.5688X70.0.0.0.8306R-squared0.    Mean dependent var10049.04Adjusted R-squared0.    S.D. dependent var12585.51S.E. of regression1041.849    Akaike info criterion16.93634Sum squared resid

9、60;   Schwarz criterion17.26329Log likelihood-247.0452    F-statistic701.4747Durbin-Watson stat2.    Prob(F-statistic)0.由此可見(jiàn)，該模型的可決系數(shù)為0.995，修正的可決系數(shù)為0.993，模型擬和很好，F(xiàn)統(tǒng)計(jì)量為701.47，模型擬和很好，回歸方程整體上顯著。但是當(dāng)=0.05時(shí)，=2.069，不僅X4、X5、X6、X7的系數(shù)t檢驗(yàn)不顯著，而且X2、X4、X6系數(shù)的符號(hào)

10、與預(yù)期相反，這表明很可能存在嚴(yán)重的多重共線性。（即除了農(nóng)業(yè)增加值、工業(yè)增加值外，其他因素對(duì)財(cái)政收入的影響都不顯著，且農(nóng)業(yè)增加值、建筑業(yè)增加值、最終消費(fèi)的回歸系數(shù)還是負(fù)數(shù)，這說(shuō)明很可能存在嚴(yán)重的多重共線性。）3、計(jì)算各解釋變量的相關(guān)系數(shù)：在Workfile窗口，選擇X2、X3、X4、X5、X6、X7數(shù)據(jù)，點(diǎn)擊“Quick”Group StatisticsCorrelationsOK,出現(xiàn)相關(guān)系數(shù)矩陣，如圖3：圖3： 相關(guān)系數(shù)矩陣X2X3X4X5X6X7X210.1470.97890.67450.46670.2465X30.14710.31880.87580.17840.6215X40.97890

11、.318810.80510.15960.4353X50.67450.87580.805110.69790.8787X60.46670.17840.15960.697910.1582X70.24650.62150.43530.87870.15821由相關(guān)系數(shù)矩陣可以看出，各解釋變量相互之間的相關(guān)系數(shù)較高，特別是農(nóng)業(yè)增加值、工業(yè)增加值、建筑業(yè)增加值、最終消費(fèi)之間，相關(guān)系數(shù)都在0.8以上。這表明模型存在著多重共線性。（三）修正多重共線性1、采用逐步回歸法，去檢驗(yàn)和解決多重共線性問(wèn)題。分別作Y對(duì)X2、X3、X4、X5、X6、X7的一元回歸，結(jié)果如下圖4：在EV主頁(yè)界面的窗口，輸入“l(fā)s y c x2”

12、，“回車鍵”。Dependent Variable: YMethod: Least SquaresDate: 10/12/10 Time: 17:49Sample: 1978 2007Included observations: 30VariableCoefficientStd. Errort-StatisticProb.  C-4086.5441463.091-2.0.0093X21.0.12.403980.0000R-squared0.    Mean dependent var10049.04Adjusted R-squared

13、0.    S.D. dependent var12585.51S.E. of regression5025.770    Akaike info criterion19.94689Sum squared resid7.07E+08    Schwarz criterion20.04030Log likelihood-297.2033    F-statistic153.8588Durbin-Watson stat0. 

14、60;  Prob(F-statistic)0.依次如上推出X3、X4、X5、X6、X7的一元回歸。綜上所述，結(jié)果如下圖4：圖4.一元回歸估計(jì)結(jié)果變量參數(shù)估計(jì)值1.0.3.0.0.0.t統(tǒng)計(jì)量12.4039828.9016822.677336.18.128950.0.0.0.0.0.0.0.0.0.0.0.-0.2、其中，加入的最大，以為基礎(chǔ)，順次加入其他變量逐步回歸。結(jié)果如下圖5：Dependent Variable: YMethod: Least SquaresDate: 10/13/10 Time: 01:27Sample: 1978 2007Included obs

15、ervations: 30VariableCoefficientStd. Errort-StatisticProb.  C1976.086388.24135.0.0000X2-1.0.-10.504860.0000X30.0.25.000560.0000R-squared0.    Mean dependent var10049.04Adjusted R-squared0.    S.D. dependent var12585.51S.E. of regression1041.474 

16、   Akaike info criterion16.82930Sum squared resid    Schwarz criterion16.96942Log likelihood-249.4395    F-statistic2103.946Durbin-Watson stat1.    Prob(F-statistic)0.依照上面，在順次加入X4、X5、X6、X7，進(jìn)行逐步回歸。綜合結(jié)果如下圖5：圖5.加入新變量的回歸結(jié)果（一）變量X2

17、X3X4X5X6X7X3,X2-1.0.0.（-10.50486）（25.00056）X3,X41.65227-9.0.（11.46367）（-8.）X3,X50.-0.0.98301（26.29703）（-5.）X3,X60.-0.0.（11.18199）（-5.）X3,X70.-0.0.（30.62427）（-2.）經(jīng)比較，新加入的方程= 0. ，改進(jìn)最大， 但是得系數(shù)為負(fù)，這顯然不符題意。在的基礎(chǔ)上分別加入其他變量后發(fā)現(xiàn)，的系數(shù)都為負(fù)，與預(yù)期估計(jì)違背。因此這些變量都會(huì)引起嚴(yán)重的多重共線性，全部剔除，只保留。修正的回歸結(jié)果為：Dependent Variable: YMethod: Lea

18、st SquaresDate: 10/12/10 Time: 17:50Sample: 1978 2007Included observations: 30VariableCoefficientStd. Errort-StatisticProb.  C-1075.289570.5337-1.0.0699X30.0.28.901680.0000R-squared0.    Mean dependent var10049.04Adjusted R-squared0.    S.D. dependen

19、t var12585.51S.E. of regression2306.678    Akaike info criterion18.38935Sum squared resid1.49E+08    Schwarz criterion18.48276Log likelihood-273.8402    F-statistic835.3074Durbin-Watson stat0.    Prob(F-statistic)0.= -10

20、75.289 + 0.（-1.） （28.90168）= 0. =0. F=835.3074這說(shuō)明在其他因素不變的情況下，工業(yè)增加值每增加1億元，財(cái)政收入平均增加0.億元。四、實(shí)踐結(jié)果報(bào)告： 為研究“農(nóng)業(yè)的發(fā)展反而會(huì)減少財(cái)政收入”的問(wèn)題，根據(jù)19782007年的財(cái)政收入，農(nóng)業(yè)增加值，工業(yè)增加值，建筑業(yè)增加值等數(shù)據(jù)，運(yùn)用EV軟件，做回歸分析，判斷是否存在多重共線性，以及修正。最后修正的回歸結(jié)果為：= -1075.289 + 0.（-1.） （28.90168）= 0. =0. F=835.3074這說(shuō)明在其他因素不變的情況下，工業(yè)增加值每增加1億元，財(cái)政收入平均增加0.億元?？蓻Q系數(shù)為0.，較高

21、，說(shuō)明模型擬合優(yōu)度高；F值為835.3074，說(shuō)明整個(gè)方程顯著；斜率系數(shù)的t值28.90168，大于t統(tǒng)計(jì)量，t檢驗(yàn)顯著，符合題意。逐步回歸后的結(jié)果雖然實(shí)現(xiàn)了減輕多重共線性的目的，但反映農(nóng)業(yè)增加值，建筑業(yè)增加值的X2,X3等也一并從模型中剔除出去了，可能會(huì)帶來(lái)設(shè)定偏誤，這是在使用逐步回歸時(shí)需要注意的問(wèn)題。附加：1、 分別作Y對(duì)X2、X3、X4、X5、X6、X7的一元回歸，結(jié)果如下：ls y c x2Dependent Variable: YMethod: Least SquaresDate: 10/12/10 Time: 17:49Sample: 1978 2007Included obser

22、vations: 30VariableCoefficientStd. Errort-StatisticProb.  C-4086.5441463.091-2.0.0093X21.0.12.403980.0000R-squared0.    Mean dependent var10049.04Adjusted R-squared0.    S.D. dependent var12585.51S.E. of regression5025.770    Akai

23、ke info criterion19.94689Sum squared resid7.07E+08    Schwarz criterion20.04030Log likelihood-297.2033    F-statistic153.8588Durbin-Watson stat0.    Prob(F-statistic)0.ls y c x3Dependent Variable: YMethod: Least SquaresDate: 10/12/10 Time:

24、17:50Sample: 1978 2007Included observations: 30VariableCoefficientStd. Errort-StatisticProb.  C-1075.289570.5337-1.0.0699X30.0.28.901680.0000R-squared0.    Mean dependent var10049.04Adjusted R-squared0.    S.D. dependent var12585.51S.E. of regression

25、2306.678    Akaike info criterion18.38935Sum squared resid1.49E+08    Schwarz criterion18.48276Log likelihood-273.8402    F-statistic835.3074Durbin-Watson stat0.    Prob(F-statistic)0.ls y c x4Dependent Variable: YMethod

26、: Least SquaresDate: 10/12/10 Time: 17:50Sample: 1978 2007Included observations: 30VariableCoefficientStd. Errort-StatisticProb.  C-1235.177727.9896-1.0.1008X43.0.22.677330.0000R-squared0.    Mean dependent var10049.04Adjusted R-squared0.    S.D. dep

27、endent var12585.51S.E. of regression2910.486    Akaike info criterion18.85437Sum squared resid2.37E+08    Schwarz criterion18.94778Log likelihood-280.8155    F-statistic514.2614Durbin-Watson stat0.    Prob(F-statistic)0.

28、ls y c x5 Dependent Variable: YMethod: Least SquaresDate: 10/12/10 Time: 17:51Sample: 1978 2007Included observations: 30VariableCoefficientStd. Errort-StatisticProb.  C-86420.4215618.35-5.0.0000X50.0.6.0.0000R-squared0.    Mean dependent var10049.04Adjusted R-squared0.&

29、#160;   S.D. dependent var12585.51S.E. of regression8310.188    Akaike info criterion20.95269Sum squared resid1.93E+09    Schwarz criterion21.04611Log likelihood-312.2904    F-statistic38.51474Durbin-Watson stat0.  

30、  Prob(F-statistic)0.ls y c x6Dependent Variable: YMethod: Least SquaresDate: 10/12/10 Time: 17:51Sample: 1978 2007Included observations: 30VariableCoefficientStd. Errort-StatisticProb.  C-2026.867934.3495-2.0.0387X60.0.18.128950.0000R-squared0.    Mean depend

31、ent var10049.04Adjusted R-squared0.    S.D. dependent var12585.51S.E. of regression3588.750    Akaike info criterion19.27334Sum squared resid3.61E+08    Schwarz criterion19.36675Log likelihood-287.1000    F-statistic328.

32、6589Durbin-Watson stat0.    Prob(F-statistic)0.ls y c x7Dependent Variable: YMethod: Least SquaresDate: 10/12/10 Time: 18:36Sample: 1978 2007Included observations: 30VariableCoefficientStd. Errort-StatisticProb.  C4934.61616135.440.0.7620X70.0.0.0.7511R-squared0. &

33、#160;  Mean dependent var10049.04Adjusted R-squared-0.    S.D. dependent var12585.51S.E. of regression12784.87    Akaike info criterion21.81425Sum squared resid4.58E+09    Schwarz criterion21.90767Log likelihood-325.2138 

34、0;  F-statistic0.Durbin-Watson stat0.    Prob(F-statistic)0.2、 以為基礎(chǔ)，順次加入其他變量逐步回歸。X3、X2:Dependent Variable: YMethod: Least SquaresDate: 10/13/10 Time: 01:27Sample: 1978 2007Included observations: 30VariableCoefficientStd. Errort-StatisticProb.  C1976.086388.241

35、35.0.0000X2-1.0.-10.504860.0000X30.0.25.000560.0000R-squared0.    Mean dependent var10049.04Adjusted R-squared0.    S.D. dependent var12585.51S.E. of regression1041.474    Akaike info criterion16.82930Sum squared resid   

36、0;Schwarz criterion16.96942Log likelihood-249.4395    F-statistic2103.946Durbin-Watson stat1.    Prob(F-statistic)0.X3、X4:Dependent Variable: YMethod: Least SquaresDate: 10/13/10 Time: 01:27Sample: 1978 2007Included observations: 30VariableCoefficientStd. Erro

37、rt-StatisticProb.  C-241.4297318.0985-0.0.4544X31.0.11.463670.0000X4-9.1.-8.0.0000R-squared0.    Mean dependent var10049.04Adjusted R-squared0.    S.D. dependent var12585.51S.E. of regression1223.617    Akaike info criterion17.151

38、65Sum squared resid    Schwarz criterion17.29177Log likelihood-254.2747    F-statistic1520.477Durbin-Watson stat1.    Prob(F-statistic)0.X3、X5:Dependent Variable: YMethod: Least SquaresDate: 10/13/10 Time: 01:28Sample: 1978 2007Included obs

39、ervations: 30VariableCoefficientStd. Errort-StatisticProb.  C27090.895304.5145.0.0000X30.0.26.297030.0000X5-0.0.-5.0.0000R-squared0.    Mean dependent var10049.04Adjusted R-squared0.    S.D. dependent var12585.51S.E. of regression1640.462    Akaike info criterion17.73798Sum squared resid    Schwarz criterion17.87810Log likelihood-263.0698    F-statistic839.9479Durbin-Watson stat0.