计量经济学一元线性回归模型实验报告.docx
计量经济学实验报告实验一：一元线性回归模型姓名:班级: 序号: 学号:实验一：一元线性回归模型一、模型设立1、问题描述：1990至2008年我国国民总收入和居民存款储蓄之间的关系。2、理论模型：Y = (3Q+舟X + 3、数据二、模型设计1、数据分析：G=20+力x20000 -I15000-°OOA 10000-0° >O O O o°5000 -Ho o o 6M1110100000 200000 300000 400000X2、参数假设：Y= 1269.686+ 0.0483X(7.0478)(37.2906)R2=0.9879F=1390.589D.W=0.21Yvs. X20000 -I15000-/10000-/Pz5000 -/1110100000 200000 300000 400000X年份国民总收入存款年底余 额年份国民总收入存款年底余 额199018718.37119.6200098000.564332.4199121826.29244.92001108068.273762.4199226937.311757.32002119095.786910.7199335260.015203.52003135174.0103617.7199448108.521518.82004159586.7119555.4199559810.529662.32005185808.6141051.0199670142.538520.82006217522.7161587.3199778060.846279.82007267763.7172534.2199883024.353407.52008316228.8217885.4199988479.259621.8Dependent Variable: Y Method: Least SquaresDate: 10/23/11 Time: 22:12Sample: 1990 2008Included observations: 19VariableCoefficientStd. Error t-StatisticProb.C1269.686180.15317.0478150.0000X0.0483270.00129637.290600.0000R-squared0.987923Mean dependent var6706.716Adjusted R-squared0.987212S.D.dependent var4078.776S.E. of regression461.2413Akaike info criterion15.20502Sum squared resid3616640.Schwarz criterion15.30443Log likelihood-142.4477F-statistic1390.589Durbin-Watson stat0.208145Prob(F-statistic)0.000000三、模型检验1、R2检验：从回归估计的结果看，模型拟合优度较好。可决系数RM.9879, 表明我国居民存款储蓄变化的98.79%可由我国国民总收入变化来解释。2、从斜率项t检验来看，大于5%显著性水平下自由度为n-2=17的临界值to.025 (17) =2.11,且有7.05>2.05； 0<0.988<1,表明我国我国国民总收入每增加I 亿元，居民存款储蓄增加0.988亿元。四、分析预测：2009年我国国民总收入为343464.7,居民存款储蓄为17174.7。则由上述回 归方程得2009年居民存款储蓄估计值为：Y=1269.7 +0.0483X343464.7=17859.0, 由此可推出相对误差为3.83%o2009 年预测区间：在样本期内 E(X)= 112506.1 Var(X)= 8867389015于是在95%置信度下E(Y2009)的预测区间为:翟只+鹄羯嘲=17859.0 + 605.3或(18464.3, 17253.7)
