《计量实习报告》word版

精品 2009-2010 学年度第 2 学期计量经济学实验报告书专 业 金融学 班 级 三班 学 号 6 学生姓名 经济与贸易学院实验一 Eviews基本操作实验一、实验目的：掌握Eviews基本操作。二、实验要求：（1）EViews软件的安装；（2）数据的输入、编辑与序列生成；（3）图形分析与描述统计分析；（4）数据文件的存贮、调用与转换。三、实验结果报告：（围绕实验要求，结合实验的内容撰写报告）一、数据的输入、序列生成 二、图形分析obsYX1985204189641986209110202198721401196319882391149281989272716909199028221854819912990216181992329726638199342553463419945127467591995603858478199669106788519978234744631998926379396obsTXX1X2198518964803532960.00011155734047319862102021040808049.80199960792e-0519873119631431133698.35910724735e-0519884149282228451846.6988210075e-0519895169092859142815.91401029038e-0519906185483440283045.39141686435e-0519917216184673379244.62577481728e-0519928266387095830443.75403558826e-05199393463411995139562.88733614367e-051994104675921864040812.13862571911e-051995115847834196764841.71004480317e-051996126788546083732251.47307947264e-051997137446355447383691.34294884708e-051998147939663037248161.25950929518e-05以上可以看出我国税收与GDP呈线性递增关系YX Mean 4309.000 35098.93 Median 3143.500 24128.00 Maximum 9263.000 79396.00 Minimum 2041.000 8964.000 Std. Dev. 2422.631 25378.06 Skewness 0.869889 0.635116 Kurtosis 2.396109 1.847265 Jarque-Bera 1.978382 1.716333 Probability 0.371877 0.423939 Observations1414实验二 一元线性回归分析过程实验一、实验目的：掌握一元线性回归模型的估计方法、检验方法和预测方法。二、实验要求：（1）会选择方程进行一元线性回归；（2）掌握一元回归分析过程；（3）掌握一元回归模型的基本检验方法；（4）会对回归方程进行经济学解释 （5）估计非线性回归模型，并进行模型比较三、实验结果报告：（围绕实验要求，结合实验的内容撰写报告）一、 图形分析 两变量趋势图分析结果显示，我国税收收入与GDP二者存在差距逐渐增大的增长趋势。相关图分析显示，我国税收收入增长与GDP密切相关，二者为非线性的曲线相关关系。 我国税收与GDP的相关图二、估计一元线性回归模型Dependent Variable: YMethod: Least SquaresDate: 06/22/10 Time: 19:29Sample: 1985 1998Included observations: 14VariableCoefficientStd. Errort-StatisticProb. C987.5417155.14306.3653640.0000GDP0.0946310.00362726.093100.0000R-squared0.982680 Mean dependent var4309.000Adjusted R-squared0.981237 S.D. dependent var2422.631S.E. of regression331.8482 Akaike info criterion14.57880Sum squared resid1321479. Schwarz criterion14.67009Log likelihood-100.0516 F-statistic680.8498Durbin-Watson stat0.796256 Prob(F-statistic)0.000000Y=987.54+0.095GDPR2=0.983 (6.37) (26.09)二、 估计非线性回归模型 1、 双对数模型Dependent Variable: LOG(Y)Method: Least SquaresDate: 06/22/10 Time: 19:45Sample: 1985 1998Included observations: 14VariableCoefficientStd. Errort-StatisticProb. C1.2704430.3316683.8304700.0024LOG(GDP)0.6822970.03241521.048660.0000R-squared0.973629 Mean dependent var8.233505Adjusted R-squared0.971431 S.D. dependent var0.528347S.E. of regression0.089302 Akaike info criterion-1.862014Sum squared resid0.095699 Schwarz criterion-1.770720Log likelihood15.03409 F-statistic443.0462Durbin-Watson stat0.476382 Prob(F-statistic)0.000000LOG（Y）=1.27+0.68LOG(GDP) R2=0.97 (3.83) (21.05)2、对数模型Dependent Variable: YMethod: Least SquaresDate: 06/22/10 Time: 19:50Sample: 1985 1998Included observations: 14VariableCoefficientStd. Errort-StatisticProb. C-26163.323149.684-8.3066490.0000LOG(GDP)2985.923307.83139.6998700.0000R-squared0.886886 Mean dependent var4309.000Adjusted R-squared0.877460 S.D. dependent var2422.631S.E. of regression848.0607 Akaike info criterion16.45535Sum squared resid8630484. Schwarz criterion16.54664Log likelihood-113.1874 F-statistic94.08748Durbin-Watson stat0.318941 Prob(F-statistic)0.000000Y=-26163.32+2985.92LOG(GDP)R2=0.887 (-8.31) (9.7)3、指数模型 Dependent Variable: LOG(Y)Method: Least SquaresDate: 06/22/10 Time: 19:55Sample: 1985 1998Included observations: 14VariableCoefficientStd. Errort-StatisticProb. C7.5086050.032400231.74630.0000GDP2.07E-057.57E-0727.268460.0000R-squared0.984118 Mean dependent var8.233505Adjusted R-squared0.982794 S.D. dependent var0.528347S.E. of regression0.069303 Akaike info criterion-2.369086Sum squared resid0.057635 Schwarz criterion-2.277792Log likelihood18.58360 F-statistic743.5689Durbin-Watson stat0.600192 Prob(F-statistic)0.000000 4、二次模型 Dependent Variable: YMethod: Least SquaresDate: 06/22/10 Time: 19:59Sample: 1985 1998Included observations: 14VariableCoefficientStd. Errort-StatisticProb. C2323.813114.422620.309040.0000GDP21.08E-064.07E-0826.652490.0000R-squared0.983388 Mean dependent var4309.000Adjusted R-squared0.982003 S.D. dependent var2422.631S.E. of regression325.0002 Akaike info criterion14.53709Sum squared resid1267502. Schwarz criterion14.62839Log likelihood-99.75965 F-statistic710.3550Durbin-Watson stat0.645855 Prob(F-statistic)0.000000四、模型比较（以二次模型、指数模型为例）二次函数回归模型残差分别表指数函数模型残差分布表实验三 多元线性回归模型一、实验目的：掌握多元线性回归模型的估计和检验方法。二、实验要求：（1）会选择方程进行多元线性回归；（2）掌握多元回归分析过程； （3）掌握多元回归模型的基本检验方法； （4）会对回归方程进行经济学解释。 （5）比较选择最佳模型三、实验结果报告：（围绕实验要求，结合实验的内容撰写报告）一、 多元线性回归模型的建立 Dependent Variable: YMethod: Least SquaresDate: 06/22/10 Time: 20:30Sample: 1978 1994Included observations: 17VariableCoefficientStd. Errort-StatisticProb. C-675.32082682.060-0.2517920.8051T77.67893115.67310.6715380.5136L0.6666650.8536260.7809800.4488K0.7764170.1044597.4327450.0000R-squared0.995764 Mean dependent var6407.249Adjusted R-squared0.994786 S.D. dependent var2486.742S.E. of regression179.5630 Akaike info criterion13.42125Sum squared resid419157.5 Schwarz criterion13.61730Log likelihood-110.0807 F-statistic1018.551Durbin-Watson stat1.510903 Prob(F-statistic)0.000000因此，我国国有独立工业企业的生产函数为： （模型1）(-0.252) (0.672) (0.781) (7.433) ，说明模型有很高的拟合优度，F检验也是高度显著的，说明职工人数L、资金K和时间变量对工业总产值的总影响是显著的。但是，模型中其他变量（包括常数项）的统计量值都较小，未通过检验。因此需要做适当的调整。 二、建立剔除时间变量的二元线性回归模型Dependent Variable: YMethod: Least SquaresDate: 06/22/10 Time: 20:36Sample: 1978 1994Included observations: 17VariableCoefficientStd. Errort-StatisticProb. C-2387.269816.8895-2.9223900.0111L1.2085320.2730204.4265280.0006K0.8344960.05742114.532870.0000R-squared0.995617 Mean dependent var6407.249Adjusted R-squared0.994990 S.D. dependent var2486.742S.E. of regression176.0069 Akaike info criterion13.33771Sum squared resid433697.8 Schwarz criterion13.48475Log likelihood-110.3705 F-statistic1589.953Durbin-Watson stat1.481994 Prob(F-statistic)0.000000此时我国国有独立工业企业的生产函数为： （模型2）(-2.922) (4.427) (14.533) 模型2的拟合优度较模型1并无多大变化，F检验也是高度显著的。但这里，解释变量、常数项的检验值都比较大，显著性概率都小于0.05，因此模型2较模型1更为合理。三、建立非线性回归模型C-D生产函数Dependent Variable: LNYMethod: Least SquaresDate: 06/22/10 Time: 20:42Sample: 1978 1994Included observations: 17VariableCoefficientStd. Errort-StatisticProb. C-1.9512531.665320-1.1716980.2609LNL0.6044670.2726972.2166250.0437LNK0.6736580.0723579.3101310.0000R-squared0.995753 Mean dependent var8.692837Adjusted R-squared0.995147 S.D. dependent var0.394921S.E. of regression0.027512 Akaike info criterion-4.189602Sum squared resid0.010597 Schwarz criterion-4.042564Log likelihood38.61162 F-statistic1641.407Durbin-Watson stat1.338201 Prob(F-statistic)0.000000C-D生产函数的估计式为： （模型3） (-1.172) (2.217) (9.310) 从模型3中看出，资本与劳动的产出弹性都是在0到1之间，模型的经济意义合理，而且拟合优度较模型2还略有提高，解释变量都通过了显著性检验。实验四 异方差模拟实验一、实验目的：了解异方差模型的检验方法和异方差模型的处理方法。二、实验要求：（1）模拟线性回归模型中随机扰动项为异方差的样本数据（2）进行Goldfeld-Quandt检验（3）利用WLS方法进行参数估计，建立模型。三、实验结果报告：（围绕实验要求，结合实验的内容撰写报告）一、人均消费与人均收入Dependent Variable: YMethod: Least SquaresDate: 06/23/10 Time: 19:15Sample: 1 27Included observations: 27VariableCoefficientStd. Errort-StatisticProb. C15.838539.4161601.6820580.1050X0.1038540.0111499.3149310.0000R-squared0.776322 Mean dependent var94.44444Adjusted R-squared0.767375 S.D. dependent var45.00712S.E. of regression21.70747 Akaike info criterion9.064377Sum squared resid11780.36 Schwarz criterion9.160365Log likelihood-120.3691 F-statistic86.76793Durbin-Watson stat2.614427 Prob(F-statistic)0.000000Y=15.84+0.104XR2=0.78 T统计 1.68 9.31 F=86.77戈德菲尔德匡特法（双变量模型）检验前1-10个数据的回归Dependent Variable: YMethod: Least SquaresDate: 06/23/10 Time: 20:18Sample: 1 10Included observations: 10VariableCoefficientStd. Errort-StatisticProb. C-3.12121010.53931-0.2961490.7747X0.1449600.0261965.5337030.0006R-squared0.792863 Mean dependent var52.50000Adjusted R-squared0.766971 S.D. dependent var20.76455S.E. of regression10.02368 Akaike info criterion7.624634Sum squared resid803.7933 Schwarz criterion7.685151Log likelihood-36.12317 F-statistic30.62187Durbin-Watson stat2.703606 Prob(F-statistic)0.000551RSS1=803.79后10个数据的回归Dependent Variable: YMethod: Least SquaresDate: 06/23/10 Time: 20:20Sample: 18 27Included observations: 10VariableCoefficientStd. Errort-StatisticProb. C48.4187056.709950.8537950.4180X0.0752110.0476311.5790270.1530R-squared0.237611 Mean dependent var136.4000Adjusted R-squared0.142312 S.D. dependent var36.04688S.E. of regression33.38354 Akaike info criterion10.03086Sum squared resid8915.686 Schwarz criterion10.09138Log likelihood-48.15430 F-statistic2.493326Durbin-Watson stat2.988119 Prob(F-statistic)0.152983 RSS2=8915.69RSS2/RSS1= 11.09F(8,8)=3.44所以存在异方差利用WLS进行异方差的消除(W=1/RESID) Dependent Variable: YMethod: Least SquaresDate: 06/23/10 Time: 19:59Sample: 1 27Included observations: 27Weighting series: RESIDVariableCoefficientStd. Errort-StatisticProb. C58.9893722.789142.5884860.0158X0.0673080.0182903.6801330.0011Weighted StatisticsR-squared0.941484 Mean dependent var-1.18E+17Adjusted R-squared0.939144 S.D. dependent var8.31E+17S.E. of regression2.05E+17 Akaike info criterion82.63371Sum squared resid1.05E+36 Schwarz criterion82.72969Log likelihood-1113.555 F-statistic13.54338Durbin-Watson stat0.338876 Prob(F-statistic)0.001121Unweighted StatisticsR-squared0.557188 Mean dependent var94.44444Adjusted R-squared0.539475 S.D. dependent var45.00712S.E. of regression30.54273 Sum squared resid23321.45Durbin-Watson stat1.287687二、 对某地区31年来居民的收入与储蓄建立的线性回归模型进行异方差检验及校正方法。Dependent Variable: YMethod: Least SquaresDate: 06/23/10 Time: 20:08Sample: 1 31Included observations: 31VariableCoefficientStd. Errort-StatisticProb. C-665.6043113.4187-5.8685560.0000X0.0845500.00468718.040560.0000R-squared0.918186 Mean dependent var1230.000Adjusted R-squared0.915365 S.D. dependent var817.1759S.E. of regression237.7341 Akaike info criterion13.84252Sum squared resid1639007. Schwarz criterion13.93504Log likelihood-212.5591 F-statistic325.4618Durbin-Watson stat1.036781 Prob(F-statistic)0.000000Y=-665.6+0.08XR2=0.918 (-5.87) (18.04)Goldfeld-Quandt检验前10个数据的回归Dependent Variable: YMethod: Least SquaresDate: 06/23/10 Time: 21:19Sample: 1 11Included observations: 11VariableCoefficientStd. Errort-StatisticProb. C-744.6351195.4108-3.8106140.0041X0.0882580.0157055.6196190.0003R-squared0.778216 Mean dependent var331.3636Adjusted R-squared0.753574 S.D. dependent var260.8157S.E. of regression129.4724 Akaike info criterion12.72778Sum squared resid150867.9 Schwarz criterion12.80012Log likelihood-68.00278 F-statistic31.58011Durbin-Watson stat1.142088 Prob(F-statistic)0.000326RSS1=150867.9后10个数据的回归Dependent Variable: YMethod: Least SquaresDate: 06/23/10 Time: 21:21Sample: 20 31Included observations: 12VariableCoefficientStd. Errort-StatisticProb. C1141.066709.84281.6074910.1390X0.0294090.0219921.3372640.2108R-squared0.151699 Mean dependent var2084.250Adjusted R-squared0.066869 S.D. dependent var287.2405S.E. of regression277.4706 Akaike info criterion14.24032Sum squared resid769899.2 Schwarz criterion14.32114Log likelihood-83.44191 F-statistic1.788274Durbin-Watson stat2.864726 Prob(F-statistic)0.210758RSS2=769899.2F=FRSS2/RSS1=5.103F(8,8)=3.44所以存在异方差利用WLS进行消除(W=1/RESID)Dependent Variable: YMethod: Least SquaresDate: 06/23/10 Time: 20:41Sample: 1 31Included observations: 31Weighting series: 1/RESIDVariableCoefficientStd. Errort-StatisticProb. C-686.076123.55233-29.129860.0000X0.0857470.00196743.582930.0000Weighted StatisticsR-squared0.995497 Mean dependent var126.3255Adjusted R-squared0.995342 S.D. dependent var1586.032S.E. of regression108.2469 Akaike info criterion12.26905Sum squared resid339804.5 Schwarz criterion12.36156Log likelihood-188.1702 F-statistic1899.471Durbin-Watson stat0.156397 Prob(F-statistic)0.000000Unweighted StatisticsR-squared0.917939 Mean dependent var1230.000Adjusted R-squared0.915110 S.D. dependent var817.1759S.E. of regression238.0918 Sum squared resid1643943.Durbin-Watson stat1.923620三、全国各地区年人均通讯费用支出与家庭可支配收入建立的线性回归模型进行异方差检验及校正方法。Goldfeld-Quandt检验前10个数据的回归Dependent Variable: YMethod: Least SquaresDate: 06/23/10 Time: 21:09Sample: 1 30Included observations: 30VariableCoefficientStd. Errort-StatisticProb. C-56.9179836.20624-1.5720490.1272X0.0580750.0064808.9620090.0000R-squared0.741501 Mean dependent var256.8727Adjusted R-squared0.732269 S.D. dependent var97.56583S.E. of regression50.48324 Akaike info criterion10.74550Sum squared resid71359.62 Schwarz criterion10.83891Log likelihood-159.1825 F-statistic80.31760Durbin-Watson stat2.008179 Prob(F-statistic)0.000000Goldfeld-Quandt检验前10个数据的回归Dependent Variable: YMethod: Least SquaresDate: 06/23/10 Time: 21:12Sample: 1 10Included observations: 10VariableCoefficientStd. Errort-StatisticProb. C-261.1499358.2945-0.7288690.4869X0.1063340.0853271.2461830.2480R-squared0.162564 Mean dependent var185.2400Adjusted R-squared0.057885 S.D. dependent var25.97864S.E. of regression25.21555 Akaike info criterion9.469655Sum squared resid5086.592 Schwarz criterion9.530172Log likelihood-45.34828 F-statistic1.552972Durbin-Watson stat3.044685 Prob(F-statistic)0.247952RSS1=5086.592后10个数据的回归Dependent Variable: YMethod: Least SquaresDate: 06/23/10 Time: 21:13Sample: 21 30Included observations: 10VariableCoefficientStd. Errort-StatisticProb. C-75.48340154.9201-0.4872410.6392X0.0604330.0216282.7941700.0234R-squared0.493907 Mean dependent var350.4440Adjusted R-squared0.430646 S.D. dependent var115.8410S.E. of regression87.40844 Akaike info criterion11.95592Sum squared resid61121.88 Schwarz criterion12.01643Log likelihood-57.77959 F-statistic7.807387Durbin-Watson stat1.846850 Prob(F-statistic)0.023407Rss2=61121.88F=Rss2/Rss1=12.02F(8,8)=3.44所以存在异方差利用WLS进行消除(W=1/RESID)Dependent Variable: YMethod: Least SquaresDate: 06/23/10 Time: 21:16Sample: 1 30Included observations: 30Weighting series: 1/RESIDVariableCoefficientStd. Errort-StatisticProb. C-46.991259.238453-5.0864850.0000X0.0562300.00171732.745880.0000Weighted StatisticsR-squared1.000000 Mean dependent var255.5239Adjusted R-squared1.000000 S.D. dependent var1400.279S.E. of regression0.025604 Akaike info criterion-4.427763Sum squared resid0.018356 Schwarz criterion-4.334350Log likelihood68.41644 F-statistic1072.292Durbin-Watson stat0.130304 Prob(F-statistic)0.000000Unweighted StatisticsR-squared0.740752 Mean dependent var256.8727Adjusted R-squared0.731494 S.D. dependent var97.56583S.E. of regression50.55628 Sum squared resid71566.25Durbin-Watson stat1.998810实验五 序列自相关模拟实验一、实验目的：了解序列相关模型的检验方法以及序列相关模型的处理方法。二、实验要求：（1）模拟线性回归模型中随机扰动项为序列自相关的样本数据，（2）进行D-W检验；（3）利用Durbin两步法进行参数估计，建立模型三、实验结果报告：（围绕实验要求，结合实验的内容撰写报告）实验六 计量经济分析的创新性实验一、实验目的：提高计量分析的创新能力。二、实验要求：（1）提出一个经济问题；（2）提出经济模型；（3）收集相关数据并进行检验；（4）建立计量经济模型，并提出对策建议。三、实验结果报告：（围绕实验要求，结合实验的内容撰写报告）.