版權說明:本文檔由用戶提供并上傳,收益歸屬內容提供方,若內容存在侵權,請進行舉報或認領
文檔簡介
1、計量經濟學上機模型分析方法總結一、隨機誤差項的異方差問題的檢驗與修正模型一:Dependent Variable: LOG(Y)Method: Least SquaresDate: 07/29/12 Time: 09:03Sample: 1 31Included observations: 31VariableCoefficientStd. Errort-StatisticProb. C1.6025280.8609781.8612880.0732LOG(X1)0.3254160.1037693.1359550.0040LOG(X2)0.5070780.04859910.4
2、33850.0000R-squared0.796506 Mean dependent var7.448704Adjusted R-squared0.781971 S.D. dependent var0.364648S.E. of regression0.170267 Akaike info criterion-0.611128Sum squared resid0.811747 Schwarz criterion-
3、0.472355Log likelihood12.47249 F-statistic54.79806Durbin-Watson stat1.964720 Prob(F-statistic)0.000000(一)異方差的檢驗1、GQ檢驗法模型二:Dependent Variable: LOG(Y)Method: Least SquaresDate: 07/29/12 Time: 09:19Sample: 1 12Included observations: 12VariableCoefficientStd
4、. Errort-StatisticProb. C3.7446261.1911133.1438040.0119LOG(X1)0.3443690.0829994.1490770.0025LOG(X2)0.1689040.1188441.4212280.1890R-squared0.669065 Mean dependent var7.239161Adjusted R-squared0.595524 S.D. dependent var0.133581S.E. of regressio
5、n0.084955 Akaike info criterion-1.881064Sum squared resid0.064957 Schwarz criterion-1.759837Log likelihood14.28638 F-statistic9.097834Durbin-Watson stat1.810822 Prob(F-statistic)0.006900模型三:Dependent Variable
6、: LOG(Y)Method: Least SquaresDate: 07/29/12 Time: 09:20Sample: 20 31Included observations: 12VariableCoefficientStd. Errort-StatisticProb. C-0.3533811.607461-0.2198380.8309LOG(X1)0.2108980.1582201.3329420.2153LOG(X2)0.8565220.1086017.8868560.0000R-squared0.878402 Me
7、an dependent var7.769851Adjusted R-squared0.851381 S.D. dependent var0.390363S.E. of regression0.150490 Akaike info criterion-0.737527Sum squared resid0.203824 Schwarz criterion-0.616301Log likelihood7.425163
8、;F-statistic32.50732Durbin-Watson stat2.123203 Prob(F-statistic)0.000076進行模型二和模型三兩次回歸,目的僅是得到出去中間7個樣本點以后前后各12個樣本點的殘差平方和RSS1和RSS2,然后用較大的RSS除以較小的RSS即可求出F統(tǒng)計量值進行顯著性檢驗。2、懷特檢驗法(White)模型一的懷特殘差檢驗結果:White Heteroskedasticity Test:F-statistic4.920995 Probability0.00
9、4339Obs*R-squared13.35705 Probability0.009657Test Equation:Dependent Variable: RESID2Method: Least SquaresDate: 05/29/13 Time: 09:04Sample: 1 31Included observations: 31VariableCoefficientStd. Errort-StatisticProb. C3.9821372.8828511.3813190.1789LOG(X1)-0.5792890.91
10、6069-0.6323640.5327(LOG(X1)20.0418390.0668660.6257100.5370LOG(X2)-0.5636560.203228-2.7735140.0101(LOG(X2)20.0402800.0138792.9021730.0075R-squared0.430873 Mean dependent var0.026185Adjusted R-squared0.343315 S.D. dependent var0.038823S.E. of regression0.0
11、31460 Akaike info criterion-3.933482Sum squared resid0.025734 Schwarz criterion-3.702194Log likelihood65.96898 F-statistic4.920995Durbin-Watson stat1.526222 Prob(F-statistic)0.004339 一方面,根據(jù)上面的Obs*R2=31*0.4308
12、73=13.357052(4),說明存在顯著的異方差問題;另一方面,根據(jù)下面的輔助回歸模型可以看出LOG(X2) 與(LOG(X2)2均通過了t檢驗,說明異方差的形式可以用LOG(X2) 與(LOG(X2)2的線性組合表示,權變量可以簡單確定為1/LOG(X2)。(二)加權最小二乘法(WLS)修正1、方法原理:具體參見教材。2、回歸結果分析模型四:Dependent Variable: LOG(Y)Method: Least SquaresDate: 07/29/12 Time: 09:06Sample: 1 31Included observations: 31Weighting serie
13、s: 1/LOG(X2)VariableCoefficientStd. Errort-StatisticProb. C1.4780850.8176101.8078110.0814LOG(X1)0.3779150.0969253.8990440.0006LOG(X2)0.4734710.0483989.7828640.0000Weighted StatisticsR-squared0.872646 Mean dependent var7.423264Adjusted R-squared0.863550
14、160; S.D. dependent var0.436598S.E. of regression0.161276 Akaike info criterion-0.719639Sum squared resid0.728274 Schwarz criterion-0.580866Log likelihood14.15440 F-statistic49.27256Durbin-Watson stat2.036239
15、 Prob(F-statistic)0.000000Unweighted StatisticsR-squared0.789709 Mean dependent var7.448704Adjusted R-squared0.774688 S.D. dependent var0.364648S.E. of regression0.173088 Sum squared resid0.838862Durbin-Watson stat2.028211加權修正
16、以后的模型四懷特檢驗結果如下:White Heteroskedasticity Test:F-statistic6.555091 Probability0.000870Obs*R-squared15.56541 Probability0.003661可以看出并沒有消除異方差性,加權修正無效。下面采用1/abs(e)權變量進行WLS回歸,結果如下:模型五:Dependent Variable: LOG(Y)Method: Least SquaresDate: 07/29/12 Time: 09:10Sam
17、ple: 1 31Included observations: 31Weighting series: 1/ABS(E)VariableCoefficientStd. Errort-StatisticProb. C1.2279290.2972684.1307080.0003LOG(X1)0.3757480.0568306.6117340.0000LOG(X2)0.5101200.01778128.688470.0000Weighted StatisticsR-squared0.999990 Mean dependent var
18、7.558578Adjusted R-squared0.999989 S.D. dependent var12.31758S.E. of regression0.041062 Akaike info criterion-3.455703Sum squared resid0.047210 Schwarz criterion-3.316930Log likelihood56.56339 F-statistic1960
19、.131Durbin-Watson stat2.487309 Prob(F-statistic)0.000000Unweighted StatisticsR-squared0.794514 Mean dependent var7.448704Adjusted R-squared0.779836 S.D. dependent var0.364648S.E. of regression0.171099 Sum squ
20、ared resid0.819694Durbin-Watson stat2.007122對加權以后的模型五進行懷特檢驗如下:White Heteroskedasticity Test:F-statistic0.199645 Probability0.936266Obs*R-squared0.923778 Probability0.921125可以看出,模型已經不再存在異方差問題,模型五可以作為修正以后的最終模型。二、隨機誤差項序列相關性問題的檢驗與修正 模型一:Dependent Variable: Y
21、Method: Least SquaresDate: 07/29/12 Time: 09:48Sample: 1991 2011Included observations: 21VariableCoefficientStd. Errort-StatisticProb. C178.975555.064213.2503050.0042X0.0200020.00113417.641570.0000R-squared0.942463 Mean dependent var922.9095Adjusted R-squared0.93943
22、5 S.D. dependent var659.3491S.E. of regression162.2653 Akaike info criterion13.10673Sum squared resid500270.3 Schwarz criterion13.20621Log likelihood-135.6207 F-statistic311.2248Durbin-Watson stat0.658849
23、0; Prob(F-statistic)0.000000 初始回歸模型一經濟意義合理,統(tǒng)計指標較為理想,但DW值偏低,模型可能存在序列相關性。(一)序列相關性的檢驗方法1、自回歸模型檢驗法Dependent Variable: EMethod: Least SquaresDate: 07/29/12 Time: 09:49Sample (adjusted): 1992 2011Included observations: 20 after adjustmentsVariableCoefficientStd. Errort-StatisticProb.
24、; E(-1)0.7170800.2018523.5524970.0021R-squared0.398929 Mean dependent var2.801737Adjusted R-squared0.398929 S.D. dependent var161.7297S.E. of regression125.3870 Akaike info criterion12.54939Sum squared resid298716.2
25、; Schwarz criterion12.59918Log likelihood-124.4939 Durbin-Watson stat1.080741說明模型一的隨機誤差項至少存在一階正序列相關性,結合該自回歸模型的DW值為1.08,懷疑存在更高階的序列相關,繼續(xù)引入e(-2)如下:Dependent Variable: EMethod: Least SquaresDate: 07/29/12 Time: 09:49Sample (adjusted): 1993 2011Included observations: 19
26、after adjustmentsVariableCoefficientStd. Errort-StatisticProb. E(-1)1.0949740.1787686.1251080.0000E(-2)-0.8150100.199977-4.0755130.0008R-squared0.692885 Mean dependent var7.790341Adjusted R-squared0.674819 S.D. dependent var164.5730S.E. of reg
27、ression93.84710 Akaike info criterion12.02051Sum squared resid149723.7 Schwarz criterion12.11993Log likelihood-112.1949 Durbin-Watson stat1.945979由于e(-2)的t檢驗顯著,說明模型一的隨機誤差項確實存在二階正序列相關性,結合該二階自回歸模型的DW值為1.95,基本確定不存在更高階的序列相關。Breusch-God
28、frey Serial Correlation LM Test:F-statistic0.888958 Probability0.431668Obs*R-squared1.998924 Probability0.368077可以看出二階自回歸模型的隨機誤差項不存在序列相關性,論證了原模型僅存在二階序列相關。2、DW檢驗法0<DW<dL 存在正自相關(趨近于0) DL<DW<dU 不能確定 DU<DW<4dU 無自相關(趨近于2)3、LM檢驗法原理:一方面,根據(jù)上面的假
29、設檢驗結果判斷是否存在序列相關性,即根據(jù)(n-p)*R2統(tǒng)計量值與卡方檢驗臨界值2(P)進行比較,其中n為原模型樣本容量,P為選擇的滯后階數(shù),R2為下面輔助回歸模型的可決系數(shù)。若(n-p)*R22(P),則拒絕不序列相關的原假設,說明模型存在顯著的序列相關性;另一方面,結合下面的輔助回歸模型中殘差滯后變量是否通過t檢驗及DW值判斷序列相關的具體階數(shù),方法與上面的自回歸模型檢驗法相同。選擇滯后一階檢驗:Breusch-Godfrey Serial Correlation LM Test:F-statistic13.15036 Probability0
30、.001931Obs*R-squared8.865308 Probability0.002906Test Equation:Dependent Variable: RESIDMethod: Least SquaresDate: 07/29/12 Time: 09:51Presample missing value lagged residuals set to zero.VariableCoefficientStd. Errort-StatisticProb. C-14.2447243.18361-0.3298640.7453
31、X0.0007140.0009070.7866170.4417RESID(-1)0.7632630.2104773.6263420.0019R-squared0.422158 Mean dependent var1.30E-13Adjusted R-squared0.357953 S.D. dependent var158.1566S.E. of regression126.7275 Akaike info criterion12.65352Sum squa
32、red resid289077.4 Schwarz criterion12.80274Log likelihood-129.8619 F-statistic6.575179Durbin-Watson stat1.159275 Prob(F-statistic)0.007183說明原模型確實存在一階序列相關性,結合該輔助回歸模型的DW值為1.16,懷疑存在更高階的序列相關,引入滯后二階檢驗如下:Breusch-Godfrey Serial Correlatio
33、n LM Test:F-statistic20.49152 Probability0.000030Obs*R-squared14.84303 Probability0.000598Test Equation:Dependent Variable: RESIDMethod: Least SquaresDate: 07/29/12 Time: 09:51Presample missing value lagged residuals set to zero.VariableCoefficientStd. E
34、rrort-StatisticProb. C14.0646332.409870.4339610.6698X-0.0006280.000742-0.8463030.4091RESID(-1)1.1084880.1761276.2936960.0000RESID(-2)-0.9181750.226004-4.0626430.0008R-squared0.706811 Mean dependent var1.30E-13Adjusted R-squared0.655072 S.D. de
35、pendent var158.1566S.E. of regression92.88633 Akaike info criterion12.07027Sum squared resid146673.8 Schwarz criterion12.26923Log likelihood-122.7379 F-statistic13.66102Durbin-Watson stat1.950263 Prob(F-stati
36、stic)0.000087由于e(-2)的t檢驗顯著,說明模型一的隨機誤差項確實存在二階正序列相關性,結合該二階自回歸模型的DW值為1.95,基本確定不存在更高階的序列相關。當然可以繼續(xù)引入滯后三階檢驗如下:Breusch-Godfrey Serial Correlation LM Test:F-statistic12.85743 Probability0.000157Obs*R-squared14.84303 Probability0.001956Test Equation:Dependent Var
37、iable: RESIDMethod: Least SquaresDate: 07/29/12 Time: 09:52Presample missing value lagged residuals set to zero.VariableCoefficientStd. Errort-StatisticProb. C14.0646733.407340.4210050.6794X-0.0006280.000765-0.8209340.4237RESID(-1)1.1082060.2713274.0844010.0009RESID(-2)-0.9175590.499523-1
38、.8368700.0849RESID(-3)-0.0006010.431119-0.0013950.9989R-squared0.706811 Mean dependent var1.30E-13Adjusted R-squared0.633514 S.D. dependent var158.1566S.E. of regression95.74504 Akaike info criterion12.16551Sum squared resid146673.
39、8 Schwarz criterion12.41421Log likelihood-122.7379 F-statistic9.643071Durbin-Watson stat1.950030 Prob(F-statistic)0.000363 可以看出并不存在三階序列相關。(二)廣義差分法修正1、方法原理參考教材自己推導二元線性回歸模型存在二階序列相關時的廣義差分模型。2、上機實現(xiàn)結果分析 模型二:Dependent Variable: YMethod:
40、Least SquaresDate: 07/29/12 Time: 09:55Sample (adjusted): 1992 2011Included observations: 20 after adjustmentsConvergence achieved after 8 iterationsVariableCoefficientStd. Errort-StatisticProb. C160.0892182.89170.8753230.3936X0.0214690.0030726.9889750.0000AR(1)0.7300780.2033523.5902230.0
41、023R-squared0.964570 Mean dependent var958.0450Adjusted R-squared0.960402 S.D. dependent var655.9980S.E. of regression130.5388 Akaike info criterion12.71870Sum squared resid289686.3 Schwarz criterion12.86806L
42、og likelihood-124.1870 F-statistic231.4107Durbin-Watson stat1.116066 Prob(F-statistic)0.000000Inverted AR Roots .73 由于AR(1)通過t檢驗,說明模型一確實至少存在一階序列相關,結合DW值為1.12,懷疑存在更高階序列相關性, LM檢驗結果如下: Breusch-Godfrey Serial Correlation LM
43、 Test:F-statistic6.380262 Probability0.009885Obs*R-squared9.193288 Probability0.010086Test Equation:Dependent Variable: RESIDMethod: Least SquaresDate: 07/29/12 Time: 09:57Presample missing value lagged residuals set to zero.VariableCoefficientStd. Error
44、t-StatisticProb. C80.86347145.26430.5566650.5860X-0.0035540.002602-1.3655560.1922AR(1)-0.5728410.437314-1.3099090.2099RESID(-1)1.0291570.3395413.0310220.0084RESID(-2)-0.1879230.598223-0.3141360.7577R-squared0.459664 Mean dependent var-7.24E-11Adjusted R-squared0.315
45、575 S.D. dependent var123.4773S.E. of regression102.1528 Akaike info criterion12.30313Sum squared resid156527.8 Schwarz criterion12.55207Log likelihood-118.0313 F-statistic3.190131Durbin-Watson stat2.021319
46、160; Prob(F-statistic)0.043963說明模型一在一階廣義差分修正后仍然存在序列相關性。繼續(xù)引入AR(2)進行修正。模型三:Dependent Variable: YMethod: Least SquaresDate: 07/29/12 Time: 09:58Sample (adjusted): 1993 2011Included observations: 19 after adjustmentsConvergence achieved after 5 iterationsVariableCoefficientStd. Errort-S
47、tatisticProb. C210.523342.671174.9336180.0002X0.0189160.00098719.173600.0000AR(1)1.0954460.1852545.9131940.0000AR(2)-0.9453840.250542-3.7733570.0018R-squared0.981385 Mean dependent var998.3158Adjusted R-squared0.977662 S.D. dependent var648.07
48、72S.E. of regression96.86089 Akaike info criterion12.16909Sum squared resid140730.5 Schwarz criterion12.36792Log likelihood-111.6064 F-statistic263.6012Durbin-Watson stat2.002336 Prob(F-statistic)0.000000Inve
49、rted AR Roots .55+.80i .55-.80i由于AR(1)和AR(2)都通過t檢驗,說明模型一確實至少存在二階序列相關,結合DW值為2.00,確定不存在更高階序列相關性,LM檢驗結果如下:Breusch-Godfrey Serial Correlation LM Test:F-statistic0.880914 Probability0.437745Obs*R-squared2.267656 Probability0.
50、321799 可以看出,二階廣義差分修正后的模型三不再存在序列相關性,可以作為最終選擇模型。三、多元線性回歸模型分析中解釋變量的選取問題多重共線性的檢驗與修正假設用解釋變量x1、x2、x3、x4來解釋Y。模型一:Dependent Variable: YMethod: Least SquaresDate: 07/29/12 Time: 10:35Sample: 1994 2011Included observations: 18VariableCoefficientStd. Errort-StatisticProb. C-43872.2714512.82-3.023002
51、0.0086X14.5610550.24699318.466320.0000X20.6704910.1300225.1567600.0001R-squared0.961029 Mean dependent var44127.11Adjusted R-squared0.955833 S.D. dependent var4409.100S.E. of regression926.6166 Akaike info criterion16.65197Sum squa
52、red resid12879274 Schwarz criterion16.80036Log likelihood-146.8677 F-statistic184.9504Durbin-Watson stat2.014913 Prob(F-statistic)0.000000模型二:Dependent Variable: YMethod: Least SquaresDate: 07/29/12 Time: 10:36Sample: 1994 2011Incl
53、uded observations: 18VariableCoefficientStd. Errort-StatisticProb. C-11978.1814072.92-0.8511510.4090X15.2559350.26859519.568280.0000X20.4084320.1219743.3485220.0048X3-0.1946090.054533-3.5686370.0031R-squared0.979593 Mean dependent var44127.11Adjusted R-squared0.9752
54、20 S.D. dependent var4409.100S.E. of regression694.0715 Akaike info criterion16.11616Sum squared resid6744293. Schwarz criterion16.31402Log likelihood-141.0454 F-statistic224.0086Durbin-Watson stat1.528658 Prob(
溫馨提示
- 1. 本站所有資源如無特殊說明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請下載最新的WinRAR軟件解壓。
- 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請聯(lián)系上傳者。文件的所有權益歸上傳用戶所有。
- 3. 本站RAR壓縮包中若帶圖紙,網頁內容里面會有圖紙預覽,若沒有圖紙預覽就沒有圖紙。
- 4. 未經權益所有人同意不得將文件中的內容挪作商業(yè)或盈利用途。
- 5. 人人文庫網僅提供信息存儲空間,僅對用戶上傳內容的表現(xiàn)方式做保護處理,對用戶上傳分享的文檔內容本身不做任何修改或編輯,并不能對任何下載內容負責。
- 6. 下載文件中如有侵權或不適當內容,請與我們聯(lián)系,我們立即糾正。
- 7. 本站不保證下載資源的準確性、安全性和完整性, 同時也不承擔用戶因使用這些下載資源對自己和他人造成任何形式的傷害或損失。
最新文檔
- 教育心理咨詢在家庭教育中的作用
- 數(shù)字化管理在小型咖啡店中的實踐應用案例分析
- 2025年度肉類產品行業(yè)技術交流與合作促進合同3篇
- 第四單元建立網站第14課一、檢查與測試網站說課稿 2023-2024學年人教版初中信息技術七年級上冊
- 第6課 現(xiàn)代科技進步與人類社會發(fā)展 說課稿-2023-2024學年高二歷史統(tǒng)編版(2019)選擇性必修2經濟與社會生活
- 7-生命最寶貴:《愛護身體 珍惜生命》(說課稿)統(tǒng)編版道德與法治三年級上冊
- 2025年房產租賃合同書6篇
- 2025年度金融行業(yè)委托招聘高級管理人員合同協(xié)議3篇
- 18《牛和鵝》第一課時(說課稿)-2024-2025學年統(tǒng)編版語文四年級上冊
- 2025年度碎石資源開發(fā)與銷售合作協(xié)議3篇
- DZ∕T 0348-2020 礦產地質勘查規(guī)范 菱鎂礦、白云巖(正式版)
- 任務型閱讀15篇(成都名校模擬)-2024年中考英語逆襲沖刺名校模擬真題速遞(四川專用)
- 高流量呼吸濕化氧療操作考核
- 2024年長春醫(yī)學高等??茖W校單招職業(yè)技能測試題庫及答案解析
- 2024年正定縣國資產控股運營集團限公司面向社會公開招聘工作人員高頻考題難、易錯點模擬試題(共500題)附帶答案詳解
- 可口可樂火炬營銷案例分析
- 赤峰市松山區(qū)王府鎮(zhèn)水泉溝礦泉水2024年度礦山地質環(huán)境治理計劃書
- 某年機關老干部工作總結
- 股骨干骨折(骨科)
- 胸心外科細化標準
- 身心靈療愈行業(yè)報告
評論
0/150
提交評論