計(jì)量經(jīng)濟(jì)學(xué)課程論文-我國公路客運(yùn)量的研究報(bào)告

1、我國公路客運(yùn)量的研究報(bào)告信息管理與信息系統(tǒng)03級指導(dǎo)老師:魯萬波白一佳 40311006師群昌40311020張斯蕊40311043陳華 40311028土一竹 40311062莊云 40311065我國公路客運(yùn)量的研究報(bào)告白一佳 陳華師群昌 王一竹張斯蕊莊云摘要：本文通過建立模型對影響我國公路客運(yùn)量的因素進(jìn)行了研究，通過Evies對七個變量進(jìn)行回歸擬合，通過建立模型 Yt =用+由Xi +p2 X2 +3 X3+P4 X 4+P 5 X5 +瓦X6 +再X 7 + Ut對樣本數(shù)據(jù)進(jìn)行回歸，分析得到最終模型 Yt = Po +Pi X 2 + P2 X 6 +P3 X 7 + ut ,并在此基

2、礎(chǔ)上細(xì)分變量優(yōu)化模型，引入虛擬變量對城市農(nóng)村的影響情況進(jìn)行對比分析，由此提出了最終模型的改進(jìn)模型Yt=Po + Pi X 2+P2 X 7 + Ut,通過樣本回歸分析得出一定的結(jié)論，提出進(jìn)一步探討的問題。關(guān)鍵詞：公路客運(yùn)量OLS回歸一.背景綜述改革開放后，我國國民經(jīng)濟(jì)持續(xù)高速發(fā)展，公路運(yùn)輸需求強(qiáng)勁增長，國家加大了公路基礎(chǔ) 設(shè)施的建設(shè)力度。隨著道路環(huán)境的改善和城鄉(xiāng)交流的日益頻繁，公路客運(yùn)量逐年提高。伴隨著 中國城市化的進(jìn)程，城鄉(xiāng)之間、城際之間的交流日益頻繁，這直接支持了公路客運(yùn)行業(yè)的發(fā)展。公路客運(yùn)在我國綜合運(yùn)輸體系客運(yùn)市場中發(fā)揮著舉足輕重的作用，承擔(dān)著90%以上的份額,因此對我國公路客運(yùn)的研究就

3、顯得很有現(xiàn)實(shí)意義，通過研究我國從改革開放至今的公路客運(yùn)量 發(fā)展變化，可以從我國國民經(jīng)濟(jì)發(fā)展的一個側(cè)面了解到我國二十多年來的交通運(yùn)輸、公共事業(yè) 建設(shè)、人民生活水平、社會生產(chǎn)、流通、分配、消費(fèi)各環(huán)節(jié)協(xié)調(diào)發(fā)展等諸多現(xiàn)實(shí)經(jīng)濟(jì)問題，對 于提升個人對國家經(jīng)濟(jì)發(fā)展認(rèn)識、研究分析的能力大有好處。因此，本文以1978年為課題研究的時(shí)間起點(diǎn)，縱觀中國公路、人口、人均收入、客運(yùn)汽車 產(chǎn)量、鐵路、民航、水路運(yùn)輸客運(yùn)量等眾多因素對我國公路客運(yùn)量的推動作用和影響，通過建 立多元線性回歸方程，進(jìn)行實(shí)證分析，得出對我國公路客運(yùn)量的顯著影響因素。二.模型變量選擇及預(yù)測在模型建立之初，我們選擇了七個對公路客運(yùn)量可能造成影響的因素

4、：客運(yùn)汽車總量、年 底總?cè)丝?、鐵路客運(yùn)量、水運(yùn)客運(yùn)量、民用航空客運(yùn)量、公路長度及全國總?cè)司杖?。從?jīng)濟(jì) 常識的角度，初步認(rèn)為，人口、人均收入作為國民經(jīng)濟(jì)衡量的基本要素對公路客運(yùn)量應(yīng)該有一 定的影響；鐵路客運(yùn)、水運(yùn)客運(yùn)、民航客運(yùn)與公路客運(yùn)存在替代的經(jīng)濟(jì)關(guān)系，其三者的客運(yùn)量 要么與公路客運(yùn)量有負(fù)相關(guān)的關(guān)系，要么與公路客運(yùn)量的相關(guān)關(guān)系不大；客運(yùn)汽車作為公路客 運(yùn)的硬件條件我們也將其引入模型，去考察客運(yùn)汽車總量與客運(yùn)量規(guī)模間的解釋關(guān)系；而客運(yùn)路線的豐富程度勢必也將對公路客運(yùn)量造成影響，在此我們用公路的長度去衡量客運(yùn)路線的豐 富程度。在以上分析的基礎(chǔ)上，進(jìn)行主觀的預(yù)測，對公路客運(yùn)量可能造成影響的因素有：

5、年底 總?cè)丝?、全國總?cè)司杖搿㈣F路客運(yùn)量、客運(yùn)汽車總量。三.模型分析根據(jù)對經(jīng)濟(jì)現(xiàn)象的分析，建立如下模型描述：Y=B0+Bl Xi +日2 X2+B3 X3 邛 4 X4 + B5 X5 + B6 X6 + B7 X 7 + u C1其中：Yt-公路客運(yùn)量,一監(jiān)運(yùn)汽車總量2 -年底總?cè)丝赬3-鐵路客運(yùn)量X4 -水運(yùn)客運(yùn)量X5 -民用航空客運(yùn)量6-公路長度X7 -全國總?cè)司杖耄ㄒ唬?、對所選擇的樣本作散點(diǎn)圖得個解釋變量與被解釋變量的關(guān)系如下系列圖所示:1500000100000050000000100200300400500X115000001000000500000090000100000110

6、000120000130000X2150000010000005000008000090000100000110000 120000X31500000100000050000001500020000250003000035000X415000001000000 -Y500000 .0,02000 400060008000 100001500000X51500000 , j 1V ,1000000.Y ,500000-,0.,80100120140160180 200X610000005000000500100015002000X7從圖形看出所選擇的解釋變量x3與x4樣本數(shù)據(jù)與所選擇的被解釋變量

7、的樣本數(shù)據(jù)間沒有x3和明顯的相關(guān)性，其余解釋變量與被解釋變量間有明顯的線性相關(guān)性。所以推測所建模型中x4對y的解釋可能不顯著。（二）、樣本模型的估計(jì)1、模型估計(jì)對所選擇的樣本數(shù)據(jù)運(yùn)用OLS法回歸得：Dependent Variable: YMethod: Least SquaresDate: 12/16/05 Time: 15:08Sample: 1 18Included observations: 18VariableCoefficientStd. Errort-StatisticProb.C-1810996.156801.2-11.549640.0000X1-18.56917178.244

8、2-0.1041780.9191X216.031731.7781879.0157720.0000X33.7978611.1424343.3243600.0077X4-2.6284404.549093-0.5777940.5762X510.8877217.879220.6089590.5561X61357.762726.40071.8691640.0911X7349.150853.140406.5703460.0001R-squared0.998779Mean dependent var941880.1Adjusted R-squared0.997924S.D. dependent var413

9、515.1S.E. of regression18842.03Akaike info criterion22.82667Sum squared resid3.55E+09Schwarz criterion23.22239Log likelihood-197.4400F-statistic1168.282Durbin-Watson stat2.666635Prob(F-statistic)0.000000(156801.2)(178.24)(1.78)(1.42)(4.55)(17.88)(726.40)(53.14)t =( -11.55)(-0.10)(9.02)(3.32)(-0.58)(

10、0.61)(1.87)(6.57)R 2 =0.9987R 2= 0.9979F =1168.28- 3.80 X 3-2.62X 4 10.88X 5 1357.76X 6 349.15X 7即:Y =1810996-18.57 X 116.03 X 2tDW =2.667從回歸的樣本模型的統(tǒng)計(jì)量R=0.998779可以看出，模型的擬合優(yōu)度非常好，從 F=1168.282x1、x4、x5參數(shù)估計(jì)可知解釋變量對模型的整體解釋顯著，然而通過樣本數(shù)據(jù)所得的解釋變量 值的t值明顯不顯著，據(jù)此推測模型解釋變量間可能存在多重共線性。2、多重共線性的檢驗(yàn)X1X2X3X4X5X11.0000000.8828

11、920.407131-0.7025490.973972X20.8828921.0000000.504735-0.5046760.920224X30.4071310.5047351.0000000.2761740.330393X4-0.702549-0.5046760.2761741.000000-0.751790X50.9739720.9202240.330393-0.7517901.000000X60.9605790.8193370.359901-0.7394020.933892X70.9076790.9248830.295472-0.7227060.974145從相關(guān)系數(shù)矩陣中可以看出，解

12、釋變量x1 與 x2、x5、x6、 x7,X6 0.960579 0.819337 0.359901 -0.7394020.933892 1.000000 0.863272X7 0.907679 0.924883 0.295472 -0.7227060.974145 0.863272 1.000000x2 與 x5、x6、x7, x5運(yùn)用相關(guān)系數(shù)矩陣檢驗(yàn)，相關(guān)系數(shù)矩陣為:與x6、x7, x6與x7高度相關(guān)，說明模型存在多重共線性。3、多重共線性的消除運(yùn)用逐步回歸法消除多重共線性: 第一步:Dependent Variable: YMethod: Least SquaresDate: 12/16

13、/05 Time: 15:25Sample: 1 18Included observations: 18VariableCoefficientStd. Errort-StatisticProb.C224417.043625.735.1441430.0001X7759.698140.5134618.751750.0000R-squared0.956478Mean dependent var941880.1Adjusted R-squared0.953758S.D. dependent var413515.1S.E. of regression88922.47Akaike info criteri

14、on25.73336Sum squared resid1.27E+11Schwarz criterion25.83229Log likelihood-229.6002F-statistic351.6280Durbin-Watson stat0.528434Prob(F-statistic)0.000000第二步：X2 x7Dependent Variable: YMethod: Least SquaresDate: 12/16/05 Time: 15:27Sample: 1 18Included observations: 18VariableCoefficientStd. Errort-St

15、atistic Prob.C-1905953.296654.0-6.4248360.0000X7X2406.146620.8351052.954207.6697712.8937677.1999910.00000.0000R-squared0.990233Mean dependent var941880.1Adjusted R-squared0.988931S.D. dependent var413515.1S.E. of regression43506.46Akaike info criterion24.35022Sum squared resid2.84E+10Schwarz criteri

16、on24.49861Log likelihood-216.1520F-statistic760.3815Durbin-Watson stat0.787593Prob(F-statistic)0.000000x2 x6 x7Dependent Variable: YMethod: Least SquaresDate: 12/16/05 Time: 15:29Sample: 1 18Included observations: 18VariableCoefficient Std. Error t-Statistic Prob.C-1956629.196460.9-9.9593800.0000X73

17、28.316939.038748.4100300.0000X62111.153468.41224.5070420.0005X219.710071.92948810.215180.0000R-squared0.996015Mean dependent var941880.1Adjusted R-squared0.995161S.D. dependent var413515.1S.E. of regression28765.19Akaike info criterion23.56485Sum squared resid1.16E+10Schwarz criterion23.76271Log lik

18、elihood-208.0836F-statistic1166.384Durbin-Watson stat1.807779Prob(F-statistic)0.000000通過加入剩余變量后剔除不顯著的變量后得:x2 x3 x6 x7Dependent Variable: YMethod: Least SquaresDate: 12/16/05 Time: 15:31Sample: 1 18Included observations: 18VariableCoefficientStd. Errort-StatisticProb.C-1877325.121383.9-15.466010.0000

19、X7393.156427.2833414.410130.0000X215.968811.40413211.372720.0000X61957.836288.53886.7853460.0000X33.2002030.6488084.9324360.0003R-squared0.998612Mean dependent var941880.1Adjusted R-squared0.998185S.D. dependent var413515.1S.E. of regression17616.05Akaike info criterion22.62114Sum squared resid4.03E

20、+09Schwarz criterion22.86847Log likelihood-198.5903F-statistic2338.575Durbin-Watson stat2.590139Prob(F-statistic)0.000000但從回歸后所得的統(tǒng)計(jì)量看，加入x3后模型的整體擬合優(yōu)度改善并不明顯，說明 x3對y的解釋能力不大；同時(shí)從經(jīng)濟(jì)意義上看，從我們先前的預(yù)測得鐵路的客運(yùn)量與公路客運(yùn)量間應(yīng)該存在負(fù)相關(guān)性，然而所估計(jì)的系數(shù)為正，與經(jīng)濟(jì)意義相違背。所以剔除x3,故最后的模型為：Yt =-1956629+328.32 X 2 2111.15 X 6 19.71X 7 (196460.9

21、) (39.04 ) ( 468.41)( 1.93)t= (9.96)(8.41)(4.51)(10.22)2- 2R =0.996015R =0.995161F =1166.384DW =1.8077794、異方差檢驗(yàn)運(yùn)用arch檢驗(yàn)得：ARCH Test:F-statisticObs*R-squared0.0002260.000256ProbabilityProbability0.9882100.987238Test Equation:Dependent Variable: RESIDA2Method: Least SquaresDate: 12/20/05 Time: 20:07Sam

22、ple(adjusted): 2 18Included observations: 17 after adjusting endpointsVariableCoefficientStd. Errort-StatisticProb.C6.00E+082.28E+082.6317560.0189RESIDA2(-1)0.0038870.2586790.0150260.9882R-squared0.000015Mean dependent var6.02E+08Adjusted R-squared-0.066651S.D. dependent var6.62E+08S.E. of regressio

23、n6.84E+08Akaike info criterion43.63541Sum squared resid7.02E+18Schwarz criterion43.73343Log likelihood-368.9009F-statistic0.000226Durbin-Watson stat1.997109Prob(F-statistic)0.988210根據(jù)F-statistic與Obs*R-squared 的P值可得模型不存在異方差。5、自相關(guān)檢驗(yàn)由DW= 1.807779，給定顯著性水平 a =0.05查表，n=18 , k=3得下臨界值和上臨界值為d l =0.933, du =1

24、.696,因?yàn)?4-1.696>1.807779>1.696, 所以模型不存在自相關(guān)性。6 .模型結(jié)論從所取樣本的估計(jì)模型得出：全國人均總收入每增加一元RMB ,其他因素不變時(shí)，公路客運(yùn)總量平均提高19.71萬人；全國總?cè)丝诿吭黾右蝗f人，其他因素不變時(shí)，公路客運(yùn)總量平均提高328.32萬人；公路總長度每增加一萬公里，其他因素不變時(shí)，公路客運(yùn)總量平均提高2111.15萬人。四.模型改進(jìn)（一）、對所選擇的樣本作散點(diǎn)圖得分類后的解釋變量與被解釋變量的關(guān)系如下系列圖所示：150000015000001000000 -1000000 -500000-500000-0,10000 20000

25、300004000050000 60000.,76000 78000 80000 82000 84000 86000 88000X21X22150000015000001000000 .1000000 .500000.500000 .0_010002000300002000400060008000 10000X72X71考慮到全國人均收入與全國總?cè)丝诖嬖趨^(qū)域差異，即可把人口范圍細(xì)分為城鎮(zhèn)和農(nóng)村。因 此，在上述模型的基礎(chǔ)上，我們進(jìn)一步考慮各細(xì)化因素的影響程度，以及農(nóng)村人口由于政策因 素而呈現(xiàn)的二次型，建立如下模型:Yt -二 0 7工1 X21 k工2 X 6 上工3 X71 - utC3Yt

26、= ：, 1 二.工2 D0 , ：1 D0 X 22 ,: 2 X 22，： 3 X 72，：4 X 6Ut其中:0 t 與1995D0 =1 t 1995X -心路客運(yùn)量21 一 ”底城鎮(zhèn)居民人口數(shù)X71 - 鎮(zhèn)居民人均收入(二)、樣本模型的估計(jì)X6X22X72公路長度-年底農(nóng)村居民人口數(shù)- -農(nóng)村居民人均收入(1)對模型。2的估計(jì) 1 .模型估計(jì) 選擇的樣本數(shù)據(jù)運(yùn)用 OLS法回歸得:Dependent Variable: YMethod: Least SquaresDate: 12/24/05 Time: 23:23Sample: 1 10Included observations: 1

27、0VariableCoefficientStd. Errort-StatisticProb.C453987.7709790.90.6396080.5461X6-11946.309828.607-1.2154620.2698X2141.153768.5584004.8085810.0030X71121.882734.826153.4997470.0128R-squared0.995571Mean dependent var641103.1Adjusted R-squared0.993356S.D. dependent var290572.5S.E. of regression23684.97Ak

28、aike info criterion23.27224Sum squared resid3.37E+09Schwarz criterion23.39328Log likelihood-112.3612F-statistic449.5276Durbin-Watson stat2.133751Prob(F-statistic)0.000000即：?Y = 453987.7+41.15376 X -11946.30 X +121.8827Xt21671(0.639608 ) (4.808581 ) ( -1.215462 ) ( 3.499747)R2 = 0.995571 F=449.5276DW

29、=2.133751上述結(jié)果，雖然方程有相當(dāng)高的擬合優(yōu)度和F值，但解釋變量的t值并不顯著，且 x6違背經(jīng)濟(jì)意義，由此推測模型的解釋變量間可能存在多重共線性。X21X71X6X2110.9681937631010.938423544908X710.96819376310110.9467378499X60.9384235449080.946737849912 .多重共線性的檢驗(yàn)：從相關(guān)矩陣可以看出解釋變量間存在高度的相關(guān)。3 .多重共線性的消除：運(yùn)用逐步回歸得到消除后的結(jié)果為：Dependent Variable: YMethod: Least Squares Date: 12/24/05 Time

30、: 23:25Sample: 1 10Included observations: 10由此得到方程:VariableCoefficientStd. Errort-StatisticProb.C-406302.355062.37-7.3789480.0002X2131.185342.52828612.334580.0000X7182.0646012.214326.7187190.0003R-squared0.994480Mean dependent var641103.1Adjusted R-squared0.992903S.D. dependent var290572.5S.E. of re

31、gression24479.23Akaike info criterion23.29236Sum squared resid4.19E+09Schwarz criterion23.38314Log likelihood-113.4618F-statistic630.5537Durbin-Watson stat1.603479Prob(F-statistic)0.000000Yt =-406302.3+31.18534 X 21 +82.06460X71(-7,378948 ) (12.33458) (6.718719)R= 0.994480F=630,5537DW=1.6034793 .異方差

32、檢驗(yàn)：Arch x21 x71ARCH Test:F-statisticObs*R-squared0.0909480.115433ProbabilityProbability0.7717380.734042Test Equation:Dependent Variable: RESIDA2Method: Least SquaresDate: 12/24/05 Time: 23:29Sample(adjusted): 2 10Included observations: 9 after adjusting endpointsVariableCoefficientStd. Errort-Statis

33、ticProb.C5.08E+082.58E+081,9700260.0895RESIDA2(-1)-0.1155490.383151-0.3015750.7717R-squared0.012826Mean dependent var4.54E+08Adjusted R-squared-0.128199S.D. dependent var5.25E+08S.E. of regression5.58E+08Akaike info criterion43,31014Sum squared resid2.18E+18Schwarz criterion43,35397Log likelihood-19

34、2.8956F-statistic0.090948Durbin-Watson stat2.056760Prob(F-statistic)0.771738可判斷模型不存在異方差。4 .自相關(guān)檢驗(yàn)：在置信度為0.1的水平下，DW=1.603479模型不存在自相關(guān)。5 .模型結(jié)論：從所取樣本的估計(jì)模型得出：城市人均總收入每增加一元RMB ,其他因素不變時(shí)，公路客運(yùn)總量平均提高82.0646萬人；城市總?cè)丝诿吭黾右蝗f人，其他因素不變時(shí)，公路客運(yùn)總量平均提高31.18534萬人。(2),對模型。3的估計(jì)1 .模型估計(jì)選擇的樣本數(shù)據(jù)運(yùn)用OLS法回歸得：Dependent Variable: YMethod

35、: Least SquaresDate: 12/25/05 Time: 22:09Sample: 1 18Included observations: 18VariableCoefficientStd. Errort-StatisticProb.C-5456506.1075174.-5.0749980.0003D05950910.3298356.1.8042050.0963D0*X22-68.0856238,50329-1.7683060.1024即:X2269.9251914.624444.7813940.0004X7277.7181728.688032.7090800.0190X61245

36、.9103307.8190.3766560.7130R-squared0.988681Mean dependent var941880.1Adjusted R-squared0.983965S.D.dependent var413515.1S.E. of regression52363.89Akaike info criterion24.83102Sum squared resid3.29E+10Schwarz criterion25.12781Log likelihood-217.4792F-statistic209.6303Durbin-Watson stat1.527338Prob(F-

37、statistic)0.000000Y=-5456506+5950910D -68.08562D X +69.92519X1245.910X +77.71817Xt002222672(-5.074998 ) (1.804205 ) (-1.768306 ) (4.781394) (0.376656)( 2.709080)R= 0.988681 F=209.6303DW=1.527338上述結(jié)果，雖然方程有較高的擬合優(yōu)度和F值，但個別解釋變量的t值并不顯著，由此推測模型的解釋變量間可能存在多重共線性。2 .多重共線性的檢驗(yàn)D0D0*X22X22X72X6D010.998996948031-0.3

38、716625096950.878283144760.790165566279D0*X22 ().9989969480311-0.3421454499550.8634536867410.764711913402X220.371662509695 -C.3421454499551-0.322713375095 -0.502528541506X720.878283144760.863453686741-0.32271337509510.9467378499X60.7901655662790.764711913402-0.5025285415060.94673784991從相關(guān)矩陣看，個別變量間存在很

39、高的相關(guān)性。3 .多重共線性的消除:通過逐步回歸得到如下結(jié)果:Dependent Variable: YMethod: Least SquaresDate: 12/25/05 Time: 21:46Sample: 1 18Included observations: 18VariableCoefficientStd. Errort-StatisticProb.C352643.347262.157.4614320.0000X72153.445010.2779014.929600.0000R-squared0.933024Mean dependent var941880.1Adjusted R-s

40、quared0.928838S.D. dependent var413515.1S.E. of regression110309.8Akaike info criterion26.16441Sum squared resid1.95E+11Schwarz criterion26.26334Log likelihood-233.4797F-statistic222.8931Durbin-Watson stat0.434010Prob(F-statistic)0.000000留x72第二步Dependent Variable: YMethod: Least SquaresDate: 12/25/0

41、5 Time: 22:01Sample: 1 18Included observations: 18VariableCoefficientStd. Errort-StatisticProb.C-2375663.484871.8-4.8995700.0002X72164.99756.35163825.977160.0000X2232.572785.7793835.6360300.0000R-squared0.978517Mean dependent var941880.1Adjusted R-squared0.975653S.D. dependent var413515.1S.E. of reg

42、ression64522.96Akaike info criterion25.13844Sum squared resid6.24E+10Schwarz criterion25.28684Log likelihood-223.2460F-statistic341.6186Durbin-Watson stat0.952903Prob(F-statistic)0.000000留 x72 x22第三步Dependent Variable: YMethod: Least SquaresDate: 12/25/05 Time: 22:03Sample: 1 18Included observations

43、: 18VariableCoefficientStd. Errort-StatisticProb.C-2414088.505274.8-4.7777730.0003D030585.2067056.750.4561090.6553X72159.912012.9194112.377650.0000X2233.111146.0544445.4688990.0001R-squared0.978832Mean dependent var941880.1Adjusted R-squared0.974296S.D. dependent var413515.1S.E. of regression66296.8

44、5Akaike info criterion25.23480Sum squared resid6.15E+10Schwarz criterion25.43266Log likelihood-223.1132F-statistic215.7906Durbin-Watson stat0.944512Prob(F-statistic)0.000000Dependent Variable: YMethod: Least SquaresDate: 12/25/05 Time:22:03Sample: 1 18Included observations: 18VariableCoefficientStd.

45、 Errort-StatisticProb.C-2396782.502043.1-4.7740550.0003X72160.913512.2894313.093650.0000X2232.886126.0029145.4783600.0001D0*X220.3041500.7749330.3924850.7006R-squared0.978751Mean dependent var941880.1Adjusted R-squared0.974198S.D. dependent var413515.1S.E. of regression66423.18Akaike info criterion2

46、5.23861Sum squared resid6.18E+10Schwarz criterion25.43647Log likelihood-223.1475F-statistic214.9529Durbin-Watson stat0.943698Prob(F-statistic)0.000000Dependent Variable: YMethod: Least SquaresDate: 12/25/05 Time:22:04Sample: 1 18Included observations: 18VariableCoefficientStd. Errort-StatisticProb.C

47、-3333059.719412.6-4.6330290.0004X72130.323721.019186.2002290.0000X2240.494927.1232655.6848820.0001X63592.7752088.1331.7205680.1073R-squared0.982267Mean dependent var941880.1Adjusted R-squared0.978467S.D. dependent var413515.1S.E. of regression60679.57Akaike info criterion25.05773Sum squared resid5.1

48、5E+10Schwarz criterion25.25559Log likelihood-221.5196F-statistic258.4966Durbin-Watson stat1.077225Prob(F-statistic)0.000000第三步的回歸中雖然各個引入的變量t值均不顯著擔(dān)任然暫留x6,繼續(xù)回歸。第四步：Dependent Variable: YMethod: Least Squares Date: 12/25/05 Time: 22:05Sample: 1 18Included observations: 18VariableCoefficientStd. Errort-S

49、tatisticProb.C-4054005.783040.4-5.1772620.0002X7289.7871730.057432.9871870.0105X2247.321317.6637196.1747180.0000X65735.0912287.4492.5072000.0262D0119447.167233.411.7766040.0990R-squared0.985731Mean dependent var941880.1Adjusted R-squared0.981341S.D. dependent var413515.1S.E. of regression56485.28Aka

50、ike info criterion24.95148Sum squared resid4.15E+10Schwarz criterion25.19881Log likelihood-219.5633F-statistic224.5224Durbin-Watson stat1.451476Prob(F-statistic)0.000000Dependent Variable: YMethod: Least SquaresDate: 12/25/05 Time: 22:06Sample: 1 18Included observations: 18VariableCoefficientStd. Errort-StatisticProb.C-4012963.778026.2-5.1578770.0002X2246.745107.5680056.1766740.0000X7290.7511330.075563.0174380.0099X65787.2962324.8222.4893510.0271D0*X221.3698020.7881681.7379580.1058R-squared0.985610Mean dependent var941880.1Adj