计量经济学作业(建立多元线性回归模型).pdf

《计量经济学作业(建立多元线性回归模型).pdf》由会员分享，可在线阅读，更多相关《计量经济学作业(建立多元线性回归模型).pdf（24页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、税收与三大产业的关系模型目录目录.”.2 l研究背景”.“.“.4 2.数据的搜集.“.5 3建立多元线性回归模型.”.”5 3 I模型估计.“.”“.”“.5 32模型检验.“”“.I O 3.2.1经济意义检验.”.”.IO 3.2 2拟合优度检验.”.”.11 3.2.3 F怆验.“.“.“.I I 3.2.4 t检验.u“”.11 3.2.5多重共线性检验.u“”.11 3.2.6白相关性检验.“.“.17 3.2.7白相关的修正.“.“.18 3.2.8异方主言。险检验.”“”.19 3.2.9异方差的修正“.“.“.23 4结论“.“.“.28 5参考文献”.u.”“.29 1 I

2、 1 1.研究背景税收是调控经济运行的重要手段。经济决定税收，税收反作用于经济。税收作为经济杠杆，通过增税与减免税等手段来影响社会成员的经济利益，引导企业、个人的经济行为，对资源配置和社会经济发展产生影响，从而达到调控宏观经济运行的目的。政府运用税收手段，既可以调节宏观经济总量，也可以调节经济结构。我国税收收入增长率在“下降”，而“质量”却在“提高”。财政部税政司发布的“2013年一季度税收收入情况分析”显示，2013年一季度全国税收总收入完成27399.20亿元，比去年同期增加2418.96亿元，增长1 0.3%.从中可以看出，一季度的税收收入增长速度改变了以往税收收入超GDP较多的增长形势

3、，呈现低速增长的态势。近年来，我国大力发展的高新技术产业、金融业、物流业三大支柱产业，成为纳税大户排行榜上最引人注目的三大集团军。这三大产业名家主专萃，在本届的三大排行表上纷纷崭露头角。因此，税收与三大产业的发展有着密不可分的联系，本文将用讨量经济学的有关方法来建立具体模型探究它们之间的具体关系。2.数据的搜集1993-2012年中国税收收入与三大产业数据统计：单位：亿元年份税收收入第一产业 第二产业 第二产业y Xl X2 X3 1993 2990.17 5342.2 9102.2 7337.099 1994 3296.91 5866.6 11699.5 9357.376 1995 4255

4、.3 6963.763 16454.43 11915.73 1996 5126.88 9572.695 22445.4 16179.76 1997 6038.04 12135.81 28679.46 19978.46 1998 6909.82 14015.39 33834.96 23326.24 1999 8234.04 14441.89 37543 26988.15 2000 9262.8 14817.63 39004.19 30580.47 2001 10682.58 14770.03 41033.58 33873.44 2002 12581.51 14944.72 45555.88 38

5、713.95 2003 15301.38 15781.27 49512.29 44361.61 2004 17636.45 16537.02 53896.77 49898.9 2005 20017.31 17381.72 62436.31 56004.73 2006 24165.68 21412.73 73904.31 64561.29 2007 28778.54 22420 87598.09 74919.28 2008 34804.35 24040 103719.5 88554.88 2009 45621.97 28627 125831.4 111351.9 2010 54223.79 33

6、702 149003.4 131340 2011 59521.59 35226 157638.8 148038 2012 73210.79 40533.6 187581.4 173087 3.建立多元线性回归模型3.1模型估计1建一个excel文挡，将数据编辘入excel文挡，进入Eviews软件包，键入file/open/foreigndata as Workfile，将excel文档导入Eviews，再进行回归分析的结果：（命令：LSY C Xl X2 X3)国We,耐le,UNillLED-巴X View I lrocl。同edlI Primtl Sa盹IDtailsl-11Show I

7、 Fe！由”。叫Delet I Gen r I Sa mp le Range:1993 201:2-2:0 obos Filter:Sample:199 3 201:2一2.0ODS 回C(ID group01 曰e sidl0 x1 囚x2:0 x3o 0 Y .！.旦旦王回Group:GROUPOl胁时，l；UNIDLED巾t阳臼XView I Proc IO创edlIP川ntI Name I Freeze 11 Defauft 2990.17 l I So比ITranspose I Edit+/-1 Smpl,obs y X1 X2 X3 1993 2990.170 5342.200 9

8、102.:200 7337099.1994 3296.910 5866.600 11699.50 9.357.376 1995 4255.300 6963.763 16454.43,11915.73 1996 5126.880 9572.695 22445.40 16179.76 1997 6038.040 12135.81 28679.46 19978.46 1998 6909.820 14015.39 33834.96 2332.6.24 1999 8234040 14441.89 37543.00 26988.15 2000 92ti2.800 14817.63 39004.19 305

9、80.47 200 10682.58 14770.03 4033.58 3.3873.44 2002 12581.51 14944.72 45555.88 38713.95 2003 15301.38 15781.27 49512.29 4436,1.61 2004 17636.45 16537.02 53896.77 49898.90 2005 20017.31 17381.72 62436.31 56004.73 2006 24165.68 21412.73 73904.31 64561.29 2007 28778.54 22420.00 87598.09 74919.28 2008 34

10、804.35 24040.00 1037司9.S.88554.88 2009 45621.97 28627.00 125自31.41113.51.9 2010 54223.79 33702.00 149003.4 1313.40.0 2011 59521.59 35226.00 157638.8 148038.0 2012 73210.79 40533.60 187581.4 173087.0 !、IE)Equation:UNID压DWorlcfile:UNIDLE。川V酬11Proc I Objed 11 Print I Name IF旧ze11 Estim曲IF。阳叫tatsI Resid

11、s I Dependent Variable:Y Method:Least Squares Date:04/16114 Time:12:19 Sample:1993 2012 Incl u oeo observations:20 Varialle C oeffi ci e nt Std.Err口rt-Statisti c Pro!.c U755.421 990 0173,1.773,121 0.095.2 II X1-0.790141 0.168211-4.6973.18 0.000:2 X:2 0.214866。087515,2.455,186 0.02:59 X3 0.354698 0.0

12、76148 4.65803:2 0.0003 R-sq,uaed0.998516 Mean cl:e pendemt var 22133.00 Adjusted R-squared 0.99823S S.D.dependent var 20863.78 S.E.of regession875.8201 AKaiK e info criteri口n16.56506 Sum squared re sid 1:2272974!。Schwacriterion16.764120 Log I ik:e Ii ho口d-U61.6505 Hannan主uinncriter.16.60393 F-statis

13、tic 3588.752 Dubin-Watson.stat 1.564939 Pro b(F-sta.tistic)0.000000 输入命令（scatXl Y）、（scatX2 Y）、（scatX3 Y）得到如下的敞点图：wi)Graph:UNTITLED W。rlcfile:UN市LED:U削”自XView I Proc I Object 11 Print!Name I Freeze 11 Options I Update 11 AddText I Line/Shade I Remo 80,000。70,000-60,000-50,000-40,000-30,000-20,000-10,

14、000-。，-ORg。10,000 20,000 30,000 40,000 50,000 X1 回Graph:UNIDLED W。耐le:UNIDLED:Untitled 巴XView I Proc IO bje ct 11 Print I Name I Fre臼11Options I Update-1 I A40 000-。10,000才。d。40,000 80,000 120.000 160.000 200,000 X2 回Graph:UNIDLED Worldile:UN市LED:Untitled飞-!:I x:View I Pmc I Obj ect 11 Print I Name

15、I Freeze 11 Options I Update I I Ad dText I Line/Shade I Remo 80.000 70.000-60,QOQ-50.000-40,QQQ-30.000-20 000-,0 俑，000才。,oo-。0 0,0。40,000 80,000 120,000 16。00.200,000 X3 估讨结果为y;=l 755.421-0.79Xli+O.215X2;+0.355X3i(-4.6973)(1.7731)R2=9985 F=3588.752 括号内为t统讨量值。3.2模型检验3.2.1经济意义检验(2.4552)DW=l.5649(4.65

16、80)我国税收收入与第一产业呈负相关，当第二产业和第三产业保持不变时，第一产业增加l个单位，我国税收收入减少0.79个单位：我国税收收入与第二产业、第三产业呈正相关，当第一产业和第三产业保持不变时，第二产业增加1个单位，我国税收收入将增加0.215个单位：当第一产业和第二产业保持不变时，第三产业增加1个单位，税收收入将增加0.355个单位。3.2.2拟合优度检验主IR2=0.9985,r-2=0.9982与l十分接近，说明其拟合优度很好。3.2.3.F检验针对比：J3,=J3 2=J3 3=0，给定显著性水平0.05在F分布表中查出自由度为3和17的临界值Foos(3,11)=3.59。由于F

17、=3588.7523.59，应拒绝原假设比，说明回归方程显著，目II第一产业（XI）、第二产业（X2）、第三产业（X3）对我国税收收入（Y）有显著影响。3.2.4 t检验分别针对比：J3j=O(j=l,2,3），给定显著性水平0.05，查t分布表得自由度为，！临界值too2s(l7)=2.110。对应 的统i-1：量分别为4.6973,2.4552,4.6580,lt,lto.阳（17)=2.11,t,不能通过显著性检验，t2，衍，to.阳（17)=2.1 l通过显著性检验。3.2.5多重共线性检验(1）税收与三大产业的由归模型：y;=l 755.421-0.79Xli+O.215X2;+0.

18、355Xs;(1.7731)R2=0.9985(-4.6973)(2.4552)(4.6580)F=3588.752 DW=l.5649 E而ti赢了UNIDLEDWorkfili:UNIDLED:,Untitled“巴XView I lroc I Object I I Print I Nam I FreeZE 11 Estimate I For,ca st I Stats I Resids DependentVariable:Y Method:Lea.st Squares Date:04/16/14 Time:12:22 Sample：才9932012 Included obse阿ation

19、s:20 Variable Coefficient c 175-5.42:1 X1 0.79014-1 X2 0.214$66 X3。.354698R-squarect 0.998516 Adjusted R-squared 099日238S.E.of regression 8758201 Sum squaeelresidl 12272974 Log liKelih口ad-161.6506 F-statistic 3588.752 Prob(Fstatistic)0 000000 Std.Error tStati stic Prob.990.0173 1.773121 0.0952。.1682

20、114.697318 0.0002 0.0$7515 2.455186 0.0259 0.076148 4.658032 0.0003 Mean dependent var 22才33.00S.D.dependent11ar 20863.78 AKai1e info criterion 16.56506 Schwarz criterion 16.76420 Hannan-Quinn criter.16.60393 Dutlin斗Natsonstat 1.S.64939 应l图表看出，可泱系数非常高，而Xl，口的t值不显著，可以推测出现该模型出现了严重的多重共线性。对所有变量进行相关性检测，得到

21、下图：固Group:UNIDLED Workfile:UNIDLEO:Untitled飞-l:l x Iv；叫ProIObject 11 lri nt Corelati。”X才X2 X3 X1 1.000000 0.991478 0.987136 A II X2 0.991478 1.000000 0.998324-.X3 0.98713;0.998324 1.000000,(2）采用逐步回归的方法来解决多重共线性的问题先对变量这个进行由归，可得X2的可决系数最大，从而得到最优简单回归方程Y=f(X2）.于是，在保留第二产业X2变量的情况下，依次加入其他变量：命令：LSYCX2IE)Equat

22、ion:UNIDLED W耐川市LED，叫“飞幡巴XV吧1叫Object11 Prin巾amej Fr,eze j IE而mateI Fore臼叫h叫ResidsI Dependent Variable:Y Method:Least Squares Date:04/16114 Time:12:241 Sample:1993 2012 Included obse阿ations:20 Variable c oeflici ent c 4525.414 X2 0398936,R-sq,uaeel0.992218 Adjusted R-squared 0.991786 S.E.of regressio

23、n 1990.951 Sum s吨ua red residl 6-436254!0 Log I ike lihood-178.22191 F-stati.stic 2295-.017 Po b(F-statistic)0000000,Std.E口!-Statistic 698.8870 6.4!75173 00083,27 47.90634 Mean dependentvar S.D.dlependent var Akaike info criterion Schwacriterion Hannan.Quinn criter.Durbin-Watson stat 回归结果为：y=-4525.4

24、14+0.3989X2(-6.4752)(47.9063)Prob.0.0000 0.0000 2,2133.00 20863.78 18.02:219 1,8 12176 1,804162 0.355-030 R2=0.9922 DW=O.3550 括号内为t统讨量值(LS Y C Xl X2 Equa6on:UNTITLED Workf;I田UNTTTLED:U ntit l ed飞呐cwlP,o。回e吭IIPdntlNomcI Freczc l IE,ti，用。tcIF。阿oo,tS.tat,J Rc,;d,I Dpnlentvaria.刷e:vMeth ocl:Le,.st Squar

25、es Date:0 411611 41 iri me:12:25 San、le:1993 202 lncluclecl obseivations:2口巳x va,1ao1 e coernc1ent Std.E.rror t-Sta创stlcProb.C川地1833.649-1.0i:;8745 0.5-984180 1 ll74.045 0.234才 170.044087 1.2.3957-4.51:;5llllll2 13.57486 0.2304 0.0003 0.0000 2233.0。20863.78 7.32205 7.4714 1 7.35t21.06795-4 R-sQuared

26、 ACJustec R-s qua.rec S.E.of re1iress，。nSun、sciua,ecresi d Log lilelinoo(F-statistic)口9965口40.996092 叮3口4.2口3:2.8916094-170.2205 2422.690 0.0000口。Mean dependent var s.D.cepencent var Akaike”o er,erion Schwa.rz criteriom Hanna.n-C注uinn criter.Durbin-Watson stat 回归结果为：y;=l833.649-1.0675X,+O.5985X2(1.2

27、440)(-4.5650)(13.5749)R 2=0.9965 DW=l.067954 括号内为t统计量值命令LS Y C X2 X3 i固M。n川LEDw叫e:U LED：川d飞偏臼xv削IProI object I IP川ntI Name I Free且11EstiYDependent Va川able:YMethod:Least Squares Date:04/16/14 Time:12:25 Sample:1993 2012 Included otservations:20 Variable Coefficient Std.Erro!-Statistic Prob.c-2379.582

28、 677.8767-3.510347 0.0027 X2-0.051561 0.099729-0.517014 0.6118 X3 0.481883 0.106498 4.524808 0.0003 R-squared 0.996470 Mean dependent var 22133.00 Adjusted R-squared 0.99605-4 S.D.dependent var 20863.78 S.E.of regression 1310.5-46 AKaiKe info criterion 17.33176 Sum squared resid 29198014 Schwarz cri

29、terion 17.48112 Log liKeliho od-170.3176 Hannan-Quinn criter.17.36091 F-statistic 2399.215 Durbin-Watson stat 0.652353 Prob(F-stati stic)0.000000 回归结果为：y=-2379.582-0.0516 x 2+0.4819 x 3(-3.5103)(-0.5170)(4.5248)R2=0.9965 F=2399.215 DW=O.6524括号内为t统i.-1量值命令：LSY C Xl X2 X3 回Equation:UNIDLE Workfile:UNI

30、DLED巾，titled飞”白XView I Proc I Object I I Print I Name I Free2e 11 Estimate I Forecast I Stats I Resids Dependle nt Variable:Y Method:Least Squares Date:04门6门41me:12:26 Sample:1993 2012 Included olse向ations:20 Variable C口efficientc 1755.421 X1-0.7争0141X2 0.214.866 X3 0.354.698 R-SQUaFed 0.998516 Adju

31、sted R-squared 0.9982,38 SJE.口fegre.ssion875.82,01 Sum square a resi a 12272:974 Log I ilee lihooel-161.6506 F-sta.tistic:J.5-88.752 Pro b(F-statistic)0.000000,StCI.IErroI-Statistic Prob.990.0173 1.77.312:1 0.0952 0.1682节 14.697319 0.0002 0.087515 2.455186 0.02:59 0.076148 4.658032 0.0003 Mean oepe

32、n cent va22133.00 S.D.depenoentvar 20863.78 Alea ilee info criterion 16.56506 Schwa.尼criterion16.764!20 Hari1a.n-Ouinn,criter.16.60393 Dubin-Watsonstat 1.5649:J.9 回归结果为：y;=l755.421-0.7901X,+O.2 149X2+0.3547X3(1.7731)(-4.6973)(2.4552)(4.6580)R2=0.9985 F=3588.752 DW=l.5649括号内为t统计量值结果分析：1）在最优简单回归方程Y=f(

33、X2）中引入变量Xl,R2由0.9922提高到0.9965，进行t检验。（Xl）不显著，可能是“多余变量”，暂时删除。2）模型中引入X3,R2没有太大变化，进行t检验自（X3）较显著。从经济理论分析，X3应该是重要因素，且X2与X3不大相关，因此可能是“有力变量飞暂时给予保留。得到如下结论：由归模型以Y=f(X2,X3）为最优模型y;=-2379.582-0.0526X2+0.4819X3(-3.5103)(-0.5170)(4.5248)R2=0.996470 F=2399.215 DW=0.6524 括号内为t统ti量值3.2.6自相关性检验(1）阁示法在OLS对话框中，键入：LSY C

34、X2 X3得下阁：回Equati。川-1:1 x View I Pmc I Object I Print I Name I Freeze 11 Estimate I Fore(ast I Stats I Resids DepenoentVariable:v Method:Least Squares Date.:04/16114 Time:12.:27 Sample:1993 2012 Included obse阿ations:20 Variable Coefficient c-2379.582 X2-0 051561 X3 0481883.R-sciu a reel 0.996470 Adju

35、sted R-sc:iuared 0.996054 S.E.of regress ion 131。.546Sum sciu a reel resicl 29198014 Log liKeli hood-170.376 F-stati stic 2399t.215,Prot(F-sta.tistic)0.000000 Std.Error!-Statistic Prob.677.8767-3.5,103.47 0.0027 0.009T29-0.5,17014 0.6118 0.106498 4.5,2.4808 0.0003 Mean depenoent11ar 22133.00 S.D.cte

36、penoentvar 20863.78 Akaike info criterion 17.3-3176 Schwarz criterion 17.48112 H annanoQu inn criter.17.3.609才Dubin-Watsonstat 0.65,2.353 在窗口中点击“View/Actual,Fitted,Residual/Actual,Fitted,Residual Graph”得到残差阁，如下：E)E quation:UNIDLED Worlcfile,UNTITLED:Untitled“白XView I Proc I Object I I Prillt I Nam e

37、 I Freeze 11 Estimate I Forecast I Stats I Resids 3,000 2.000 1.000。-1,000-2.000 133喝惜961=2000 2002.2004 2006.2008 2010 2012 Rosidu副一A出到一FittEd I 80,000 回，0004-0,000 20,000。主l图可知，残差的舟，歹1J图是循环型的，已不是频繁改变符号，而是连续几个正值后连续几个负值再又连续几个正值，表明存在正相关。(2)DW检验根据前阁伯i才的结果DW=0.6524，给定显著性水平0.05，查DW袭，因为T=20，解释变量的个数k为2，得下

38、限临界值d1=l.20，上良临界值du=l.41。因为ODW=O.6524d1=l.20，表明存在正自相关。3.2.7自相关的修正广义差分法：在命令行中输入：LSYl C X2 X3 AR(l），得下阁：El Equatio川懦巴XView I Proc I Object I I Print I N 1am e I Freeze I I Esti肌ateI Foreast I Stat叫ResidsDependent Variable：Method:Least:Squares Date:04/16/14 Time:12:28 Sampleadjusteel)1991 4 201.2 Inclu

39、ded obsevati口ns:19 aner aeljustm e nts convergence ach ieveo aner 1 o iterations Variable c oeffi ci e nt Std.Error t-Statistic c-79410.934!35-59.63,7-2.230827 X2 017660 0.058965 2.996507 X3 0.276277 0.067297 4.105323 AR(1)0.840512 0.1125-2.6 7.4!69510 R-sciuaeel0.999201 Mean dependent11ar Adjusted

40、R-squared 0.999041.S.D.c!ependent var S.E.of regession648.0304!Akaike info criterion Sum sciu a red res id 6299151.Schwa尼criterionLog likelihood-147.7190 Hanna1-0uinn criter.F-stati stic 6253.711 1 D u rbin-Watson stat Pro b(F-.sta.ti sti c)0.000000 Inverted AR Roots.84 Prob.0.0414 0.0090 0.0009。000

41、02.3140.51 2.0929.66 15.97042 16.16925 16 00407 1.7887,84 其中，DW=l.7888，和之前的DW=0.6524比起来有了较大提高，且给定显着性水平 0.05时，d.=l.4l DW=O,652440,000 。30,000。20,000 。10,000 。J。40,000 80,000 120,000 160,000 200,000 X2 日Graph:UN可LEOWorkfile:UNffiLEO:Untitled飞仰自XView ProObject Print Na me Freeze Options Update AddTel4-

42、0,000-。10,000-J,sf！。，、0。40,000 80,000 120,000 160,000 200,000 X3 佳散点阁可知，随着我国税收收入的增加，第二、三产业产值水平不断提高，但离散度也随之扩大。这说明税收第二、三产业间存在递增的异方差性。在多重共线性检验时己确定了最优模型Y=f(X2,X3），确认我国税收收入第一产业相关关系不明显。(2)llhi te检验建立自归模型：LSY X2口，回归结果如下阁：国Equation:UNillLED W。rkf阳UN市LED：圳，tied“Clx View I Proc I Object 11 Print I Name I Fre叫

43、IEstimate I Fore臼stI Stats I陆仙Dependent Variable:Y Method:Least Squares Date:04/16/14 Time:12:30 Sample:1993 2012 Included obseivations:20 Variable Coefficient X2-0.307592 X3 0.743422 R-squared 0.993911 Adjusted R-squared 0.993572 S.E.of regression 1672.695 Sum squared resid 50362340 Log likelihood-

44、175.7690 Durbin-Watson stat 0.912584 Std.Error t-Statistic Prob.0.086811-3.543239 0.0023 0.097126 7.654226 0.0000 Mean dependent var 22133.00 S.D.dependent var 20863.78 Akaike info criterion 17.77690 Schwarz criterion 17.87647 Hannan-Quinn criter.17.79634 在方程窗口上点击“viewresidualdiagnost icsHeteroskeda

45、sticity Tests”，选择white，检验结果如下所示：国Eq州。町UNillLEDWorkf阳UNillLED:Untitled 懦OX View I Proc I Object 11 Point I Name I Freeze I I Estimate I Forecast I Stats I Res ids Heteroskedasticity Test:White F-statistic ObsR-squared Scaled explained SS 3.028766 Prob.F(3,16)7.244!033 Prob.Chi-Square(3)3.148659 Prob.

46、Ghi-Square(3)Test Equation:Dependent Variable:RESIDA 2 Method:Least Squares Date:04/16/14 Time:12:J.1 Sa盯1ple:1993 2012 Included observation.s:20 Variable Coefficient c 3197630.X2A 2-0.062128 X2X3 0.136495 X3A2-0.074!864 R-squared 0.362202 Adjusted R-squared 0.242614 S.E.of regression 2329259.Su1 sq

47、uared resid 8.68E13 Log likelihood-319.3685 F-statistic 3.028766 Pro b(F-statistic)0.060009 Std.Error I-Statistic 884537.5 3.615030 0 027670-2.245265 0.061336 2.225375 0.034018-2.200738 Mean dependent var S.D.dependent var AKaike info criterion Schwarz criterion Hannan-Quinn criter.Durbin-Watson sta

48、t 0.0600 0.0645 0.3693 Prob.0.0023 0.0392 0.04!08 0.04!28 2518117.2676450.32.33685 32.53600 32.37573 0.644348 其中F值为辅助回归模型的F统剖量值。取显著水平 0.05，由于os(5)=11.07nR2=lO.5426，所以不存在异方差。性。3.2.9异方差的修正(1）确定权数变量取以下三种形式作为权数变量：Wl=l/x,v w2=1/I e,I w3=1/ec 生成权数变量：GENRWl=l/X2向 0.5 GENR W2=1/X3-0.5 GENR W3=1/ABS(RESID)GE

49、NR W4=1/RESID民2(2）利用加权最小二乘法估i.-1模型在Eviews命令窗口中键入：LS(W=Wl)Y C X2 X3得下阁：国Egu础。n:UNTTLED Wor时，le:UNTITLED:Untitled飞v;e切IProc I Obj ect I I Print I Name I Freeze 11 Est imate I Foreca叫StatsI Resids c:l x Depenctent Variable:Y Method:Least SQuar-es Date:04116114 Time:12.,3,3 Sa1ple:1993 2012 In cl u,die

50、di observations:20 leighting1 series:W1 leiglilt type:lnvesestandard deviation lEView s default scaling)Variable Coefficient Proo.Std Eriror t-Stati sti c C尼泊叫4!80.57960.235566 0.661083 494,34 16,0.099668 Ql.108940-0.972161-2,363,500 6.1068300 0.3,4!46 0.0303 0.0000 甲VeightedStatistics R-sq,uared Ad