计量经济学实验报告(一元线性相关).doc

计量经济学实验报告一元线性回归模型上善若水中国1980-2007年全社会固定资产投资总额与工业总产值的关系一. 模型设定1. 问题描述：中国1980〜2007年全社会固定资产投资总额与工业总产值的关系2. 理论建模 Y二优+0]X+〃3. 数据中国1980〜2007年全社会固定资产投资总额X与工业总产值Y的统计资料如下表 所示4.单位：亿元年份全社会固定资产 投资(X)工业增加值(Y)年份全社会固定资产 投资(X)工业增加值(Y)1980910.91996.5199417042.119480.719819612048.4199520019.324950.619821230.421623199622913.529447.619831430」2375.6199724941.132921.419841832.92789.0199828406.234018.419852543.23448.7199929854.735861.519863120.63967.0200032917.740033.619873791.74585.8200137213.543580.619884753.85777.2200243499.947431.319894410.46484.0200355566.654945.5199045176858.0200470477.465210.019915594.58087.1200588773.677230.819928080.110284.52006109998.291310.9199313072.314188.02007137323.9107367.2二. 模型估计Y2.参数估计y= 5179.343 +0.8176X(30.54) (4.359)R2=0.9729 F=932.7681 D.W=0.16Dependent Variable: YMethod: Least SquaresDate: 10/26/11 Time: 23:05Sample: 1980 2007In eluded observati ons: 28VariableCoefficientStd. Error t-StatisticProb.X0.8176260.026771 30.541250.0000C5179.3431188.223 4.3588990.0002R-squared0.972882Mean dependent var27815.79Adjusted R-squared0.971839SD. dependent var29284.74S 匚.of regression4914.354Akaike info criterion19.90646Sum squared resid6.28EP8Schwarz criterion20.00161Log likelihood-276.6904F-statistic932.7681Durbin-Watson stat0.157154Prob(F-statistic)0.000000三. 模型检验1. R?检验：从回归估计的结果看，模型的拟合优度较好。

可决系数R2=0.9729,表明 我国工业增加值变化的99.74%可由我国全社会固定资产投资总额变化來解释2. t检验:大于5%显著性水平下自由度为n-2=28的临界值t0.025 (28) =2.048,且有 30.54< 2.048； 0<0.818<1,表明我国全社会固定资产投资增加1亿元，工业增 加值增加0.817亿元四.分析预测XYMean27685.5927815.79Media n15057.2016834.35Maximum137323.9107367.2Minimum910.90001996.500Std. Dev.35327.8329284.74Skewness1.7208791.196631Kurtosis5.2568063.604412Jarque-Bera19.762017.108518Probability0.0000510.028603Sum775196.6778842.2Sum Sq. Dev・3.37E+102.32E+10Observati ons2828由图表1可看出，2008年社会固定资产总投资预计为160000利用所估计模型可预测 这时的工业增加值为 Y=5179.343+160000 X0.8176=135995.343 □取a=0.5,则2008年工业增加值Yf平均值置信度为95%的预测区问为135995.343±2.048X4919.354X V 1/30+ (160000-27685.59) A2/37441667200或(128864.682,143126.004)取a=0.5,则2008年工业增加值Yf个别值置信度为95%的预测区间为135995.343 ±2.048 X 4919.354 X V 1 + 1/30+ (160000-27685.59) A2/37441667200或(123652.382,148338.304)Hiew | Proc | Object | Print | Name | Freeze | Estimate | Forecast | Stats | Resids |YFForecast: YFActual: YForecast sample: 1980 2007 Included observations: 28Root Mean Squared Error 4735.590Mean Absolute Error 4134.202Mean Abs. Percent Error 53.71801Theil Inequality Coefficient 0.059391Bias Proportion 0.000000Variance Proportion 0.006873Covariance Proportion 0.993127。