序列相关,虚拟变量,多重共线

《序列相关,虚拟变量,多重共线》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。

1、1．序列相关实验：下表是1965年-1994年美国人均真实工资Y与人均产出指数X的数据，通过建立模型进行分析：年份YX196569.358.6196671.861196773.762.3196876.564.5196977.664.919707966.2197180.568.8197282.971197384.773.1197483.772.2197584.574.819768777.2197788.178.4197889.779.519799079.7198089.779.8198189.881.4198291.181.2198391.284198491.586.4

2、198592.888.1198695.990.7198796.391.3198897.392.4198995.893.3199096.494.5199197.495.91992100100199399.9100.1199499.7101.4（1）用普通最小二乘法估计模型参数DependentVariable:YMethod:LeastSquaresDate:05/25/14Time:14:48Sample:19651994Includedobservations:30VariableCoefficientStd.Errort-StatisticProb.C33.429

3、562.04155316.374570.0000X0.6801150.02509227.105040.0000R-squared0.963288Meandependentvar88.12667AdjustedR-squared0.961976S.D.dependentvar8.691372S.E.ofregression1.694788Akaikeinfocriterion3.957333Sumsquaredresid80.42455Schwarzcriterion4.050746Loglikelihood-57.35999F-statistic734.6830Dur

4、bin-Watsonstat0.297779Prob(F-statistic)0.000000（1）用图形法判断模型是否存在一阶自相关（2）用DW值检验模型是否存在一阶自相关对于样本容量30，查DW便可知=1.35,=1.49，模型中DW<，所以模型存在自相关，这一点可以从第二题图中看出来（3）使用迭代法估计模型参数，并证明在5%显著性水平下迭代后的模型随机扰动项不存在一阶自相关。DependentVariable:YMethod:LeastSquaresDate:05/25/14Time:14:48Sample:19651994Includedobservation

5、s:30VariableCoefficientStd.Errort-StatisticProb.C33.429562.04155316.374570.0000X0.6801150.02509227.105040.0000R-squared0.963288Meandependentvar88.12667AdjustedR-squared0.961976S.D.dependentvar8.691372S.E.ofregression1.694788Akaikeinfocriterion3.957333Sumsquaredresid80.42455Schwarzcriter

6、ion4.050746Loglikelihood-57.35999F-statistic734.6830Durbin-Watsonstat0.297779Prob(F-statistic)0.0000002．虚拟变量实验：下表给出了1965-1970年美国制造业利润和销售额的季度数据。假定利润不仅与销售额有关，而且和季度因素有关。要求对下列情况分别估计利润模型：利润销售额YX1965-I10503114862II12192123968III10834121454IV122011319171966-I12245129911II14101140976III12213137

7、828IV128201456451967-I11349136989II12715145126III11014141536IV127301517761968-I12539148826II14949158913III13203155727IV149471684091969-I14151162781II16049176057III14024172419IV143151833271970-I12381170415II14091181313III12174176712IV10985180370（1）如果认为季度影响使利润平均值发生变异，应如何引入虚拟变量？哪个