一元线性回归模型算例

400Р26500Р1829Р26760Р2200Р28300Р2017Р27430Р2105Р29560Р1600Р28150Р2250Р32100Р2420Р32500Р2570Р35250Р1720Р33500Р1900Р36000Р2100Р36200Р2800Р38200Р根据经济理论建立计量经济模型Р Р用普通最小二乘法进行估计,结果如下РDependent Variable: YРMethod: Least SquaresРDate: 08/27/07 Time: 21:51РSample: 1 31РIncluded observations: 31РVariableРCoefficientРStd. ErrorРt-StatisticРProb. РCР-700.4110Р116.6679Р-6.003458Р0.0000РXР0.087831Р0.004827Р18.19575Р0.0000РR-squaredР0.919464Р Mean dependent varР1266.452РAdjusted R-squaredР0.916686Р S.D. dependent varР846.7570РS.E. of regressionР244.4088Р Akaike info criterionР13.89790РSum squared residР1732334.Р Schwarz criterionР13.99042РLog likelihoodР-213.4175Р F-statisticР331.0852РDurbin-Watson statР1.089829Р Prob(F-statistic)Р0.000000Р利用上述结果计算残差,观察残差的取值,好像随子变量的变化而变化,怀疑存在模型异方差性,用图示法检验,即绘制自变量与残差的散点图如下