我国国内生产总值的多元线性回归分析

由表3检验结果表明,含一阶滞后变量时的,故随机扰动项存在一阶序列相关性;含二阶滞后变量时的,故随机扰动项不存在二阶序列相关性。4.2.3序列相关性修正采用科克伦-奥科特迭代法,得表4所示结果:表4科克伦-奥科特迭代回归结果DependentVariable:YMethod:LeastSquaresDate:12/26/02Time:16:38Sample(adjusted):19842011Includedobservations:28afteradjustmentsConvergenceachievedafter15iterationsVariableCoefficientStd.Errort-StatisticProb.  C-6851.2622542.199-2.6950140.0136X10.3301150.0563585.8574630.0000X22.5872010.18742913.803630.0000X390.243369.6984469.3049300.0000X40.2110950.0398955.2913150.0000AR(1)1.0380940.2197254.7245220.0001AR(2)-0.5695740.216848-2.6266030.0158R-squared0.999770    Meandependentvar122430.4AdjustedR-squared0.999704    S.D.dependentvar128865.4S.E.ofregression2217.613    Akaikeinfocriterion18.45857Sumsquaredresid1.03E+08    Schwarzcriterion18.79162Loglikelihood-251.4200    F-statistic15191.96