第三章13题下表列出了中国某年按行业分的全部制造业国有企业及规模以上制造业非国有企业的工业总产值Y,资产合计K及职工人数L。
解: ⑴ 先对丫二AK〉L:e .左右两边同时取对数得:In 丫二 C : In K 1 In L C = In A": _d n e相应的数据变为:LOGr LOGL LOGK8.222204 4 7273B3 8,0321077274147 4.204593 7.4291837.468724+.4308177 91S7247.280208 3295837 7 507726&.546616579G9600.5655877.736014 4.737+92 7,4723707204276 4.060443 6.8449226J87334 3433987 5.5438265.S139S9 2 772569 5 8957247.371716 41SS655 7.8283316.42439g 4.060443 0,8811346.426391 3.332205 6.2i612€8.395972 41108748.2390428.6S6785 5537334 9.&697017.485138 4 41B841 7 9359327.125339 3.4965097.5002206.700362 37612007.0200217 549451 4.110374 7.526^468.214154 5480639 87191918.462293 54D2677 91300257.697196 4 382027 7.9&08997.839825 4.664548 7 3421333.021896 5.402577 8,4738477692S5T 5.093750 8.083037通过Eviews软件进行回归分析得到如下结果:Dependent Variable: LOGYMeihod: Least SquaresDa-te:03?25V14 rime- 00:33Sarnie: 1 24Included obseivations: 24Variable Coefficient Std. Error l-Statisbc Pirob.LOCK0.7251310,127391 5.6921700.0000LOGL馭29653 0.132647 0.97742503395c 1 352738 0.561254 2.410253 0 3252R*squared 0.092396 Mean dependent vsr 7507&37Adjusted 弘squared 0.882139 S D d&pendent-.ar 0.7385&8S.E. o1 regression0.253554 Aka ike into critencn 0.209&88Sum squared resid 1 350031 SchVi arz trite riqn 0.357244Log livelihood 0 480149 Hannan-Quinn criter. 0.249056^statistic87.07231Dirbin-Watson std1,387020PrQt»(F-siati siic} 0OCOCOO于是得到回归方程为:C曾iniM也id:口UOGY DDGK LOGL CEEtimatifin Eq.u*ti"M W ■1J0GT = COLOGK + CQ)*L J OGL + CG)E心応Coefficients;UKT = □ 72513051 34E3fLOSE + 0 1ES653O32923*WGL t 1.352730254072 .,首先可决系数R =0.892388和修正的可决系数R =0.882139都是接近于1的,故该回归方程的模拟情况还是比较好的。
在5%的显著性水平下,自由度为(2, 21的F分布的临界值为F o. o52, 21> 3. 47该回归分析的统计量F =87.07231显著大于3.47,因此In Y与InK、lnL有显著的关系;再看t分布,因为t o.o(s2 1> 1. 721其常数项庐0 = 2,41 0253 1.、72h1K的系数■^5. 6921 70 1. 7说明这两项已经通过检验,但是InL的回归系数没有通过检验。
⑵ 这个题不知道怎么做,只能根据答案提示做出结果,具体不知道怎么分析Wald TestEquation: Untitled第四章8题下表列出了某年中国部分省市城镇居民家庭平均每个全年可支配收入(支出(YX)与消费性的统计数据。
解:⑴ 最小线性二乘估计的检验结果和回归方程为:on Conmand:LS Y I CEsti.m血E^u.^11 >n.Y = C(1)^X + C(2)Subst i tultd eikl.s.I = 0.SO5975597O4M + 13. 261042^541Vkew Prot]obJ«t| Estim酬屮眈cast StatslRrfsi d$Deperdent Variable; Yf.leth&d: Least SquaresDate:03J25J14 Time 01:26Sample-1 16IMluded observations: 16Variable Coefficient Gtd. Error t-Statistic Prob.X 0805976 0(328103 28.^7838 0.0000C 13.26104 175 9228 0.075537 0 9409R-squared 0.983263 Mean dependent var 4859 794Adjusted R-squared 0.98206S SB. dependentuar1449.C29S.E, of regression 154 1222 AKaiKe info criterion 13.49132Sum squared resid 5275562 Schwarz criterion 13.58790Leg likelihood -105.9306 Hannan-Quinn enter 13.49627F-slatislic 922.477S Durbin-Watson stat 1.M7407Prob(F-statistic>0.000000⑵异方差检验X与丫散点图,从下图可以看出方差基本一致怀特检验结果:(这个表有些项看不懂,故也不知道怎么分析)F-statisiic 1.039744 Prob.F(1t14) 0.3252Obs"R-squared 1.106129 Prob. Chi-Square(i) 0.2929Scaled explained SS 0.855881 Prob. Chi-Square(l) 0.3549Test Equaten:Dependent Variable: RESID A2Mettiad: Least SqjaresDate. 03/25/-T4 Time. 0d 43Sample: 1 1SIncluded observations: 16/aiable Coefficient Std Error t-Statistic Prob.C 14175.05 22C4^ 7S 0.6430120.5305X A2 0.000480 0.000+71 1.019678 0 3252 R-squared 0.0&9133 M&an deoeridentvar 32973.01Adjusted R-squared 0.002S43 S O. dependents 46415 41S.E. of regression 43351 39 Akaike info criterion 24.52535Suim squared resid 3.27E^10 Schwarz criterion 24.62342Log likelihood -194 2148 Hannan-Quinn criter. 24.53179F-shtistic 1.030744Durbin-Watson vtat 2.178440Prob(F-staisflc) 0325179G-Q检验:两个样本的估计结果为:De pendent Variable: Y Methiod: Least Squares Date. 03/25/14 Time: 02:29Sample: 1 &Included observations: 6Va ri^ble CoefTicient Std. Enor t-Slatis:tic Prot).X 0.8S6823 0.416049 2155569 0.Q974c -369.0670 2029.240 -0 1S1975 0.8645R-squared0.537384l.lean deperdent var4003.505Ac^Listmci R-squared 0 421730 S.D. dependent var176.5006SE. of regression 134.2182Akaike info criterion12.39301Sum squared r&sid 72058.15 Sctiwarz criterion 12.S29fiOLog likelihood -3&«9404 Hannan-Quinn criter 12.&2014^-statistic 4.646479 Durbin-Watson stai 3039775Pro&(F-statistic) 0 D57371Dependent Va riatlei YMethod: Le^st SquaresDate: 03/25/14 Time; 0229Sample: 1 6Included observations. Sh ariable Coefficient Std 匚r「or 1-Statistic ProbX 0.84&336 0X57921 1461199 0.0001C -353.4S40 455 613S 0.775798 04012R-squared 0.931610 Mean d&pandentvar6103.605Adjusted R-sq liar Ed0.977013 S □ dependent var1792242S E. of regression 271Y325 AKaite inro 匚riterion 14 30872Sum squared resid 2953543 Sclwarz criterion 1423930Log likelinaod -40.92615 Hannan-Quinn criter. 14 03085F-statistic213.5104 Durtjin</a1son stat 3.270444Pro b(F-stati stic)0 000128RSS“ = 295354.3 =4.09883268 £F0o5(4,4) =6.39,故接受原假于是得到如下的F统计量:F =RS^/4 72058.15设,即不是异方差的。