《序列相关,虚拟变量,多重共线.doc》由会员分享，可在线阅读，更多相关《序列相关,虚拟变量,多重共线.doc（6页珍藏版）》请在金锄头文库上搜索。
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YX196569.358.6196671.861196773.762.3196876.564.5196977.664.919707966.2197180.568.8197282.971197384.773.1197483.772.2197584.574.819768777.2197788.178.4197889.779.519799079.7198089.779.8198189.881.4198291.181.2198391.284198491.586.4198592.888.119
2、8695.990.7198796.391.3198897.392.4198995.893.3199096.494.5199197.495.91992100100199399.9100.1199499.7101.4（1） 用普通最小二乘法估计模型参数Dependent Variable: YMethod: Least SquaresDate: 05/25/14 Time: 14:48Sample: 1965 1994Included observations: 30VariableCoefficientStd. Errort-StatisticProb.C33.429562.04155316.3
3、74570.0000X0.6801150.02509227.105040.0000R-squared0.963288Mean dependent var88.12667Adjusted R-squared0.961976S.D. dependent var8.691372S.E. of regression1.694788Akaike info criterion3.957333Sum squared resid80.42455Schwarz criterion4.050746Log likelihood-57.35999F-statistic734.6830Durbin-Watson sta
4、t0.297779Prob(F-statistic)0.000000（2） 用图形法判断模型是否存在一阶自相关（3） 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（4） 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。Dependent Variable: YMethod: Least SquaresDate: 05/25/14 Time: 14:48Sample: 1965 1994Included observations: 30Variabl
5、eCoefficientStd. Errort-StatisticProb.C33.429562.04155316.374570.0000X0.6801150.02509227.105040.0000R-squared0.963288Mean dependent var88.12667Adjusted R-squared0.961976S.D. dependent var8.691372S.E. of regression1.694788Akaike info criterion3.957333Sum squared resid80.42455Schwarz criterion4.050746
6、Log likelihood-57.35999F-statistic734.6830Durbin-Watson stat0.297779Prob(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-I12245129911II14101140976III12213137828IV128201456451967-I1
7、1349136989II12715145126III11014141536IV127301517761968-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？哪个季度对利润平均值有显著影响（显著性水平为0.1）？你怎么得出这个结论的？并写出该季度的利润
8、平均值。Dependent Variable: YMethod: Least SquaresDate: 05/25/14 Time: 15:03Sample: 1965:1 1970:4Included observations: 23Excluded observations: 1VariableCoefficientStd. Errort-StatisticProb.C6609.5912126.0463.1088660.0061X0.0394790.0124593.1688470.0053D1-98.50491727.3934-0.1354220.8938D21311.300689.167
9、11.9027310.0732D3-325.1485699.3650-0.4649200.6476R-squared0.536543Mean dependent var12892.35Adjusted R-squared0.433552S.D. dependent var1479.083S.E. of regression1113.198Akaike info criterion17.05752Sum squared resid22305772Schwarz criterion17.30437Log likelihood-191.1615F-statistic5.209634Durbin-Wa
10、tson stat0.402407Prob(F-statistic)0.005750第二季度对利润平均值有显著影响，根据t检验可以得出，该季度的利润平均值是4938.30637。（2）如果认为季度影响同时使利润平均值和利润对销售额的变化率发生变异，应如何引入虚拟变量？3多重共线性实验：设被解释变量Y与解释变量X1、X2、X3、X4、X5、X6的数据如下：年份YX1X2X3X4X5X619887.459.12425.517.517.8185.8521.6819897.6059.32422.322.919.51185.3521.0819907.8559.3641823.718.93185.121.
11、0319917.8059.2419.221.119.05184.820.7319926.99.86384.223.319.57184.621.9319937.478.7372.519.119.95184.2522.4919947.3859.46372.918.220.89181.3523.2619957.2259.88380.822.223.27179.324.3919968.1310401.727.626.06178.125.0419978.729.8406.528.828.55176.2525.5319989.14510.26410.527.830.12174.3526.64199910.
12、1059.6244724.432.78174.2527.53200010.179.44452.824.132.21179.3528.12200110.5410.66467.127.833.57173.8531.35200210.63510.68495.219.534.86179.534.58200310.45512.3250025.436.6166.8541.78200410.99513.352528.440.35158.2542.85（1）将Y关于其他变量线性回归，对方程进行检验，检验结果说明什么？Dependent Variable: YMethod: Least SquaresDate:
13、 05/25/14 Time: 15:09Sample: 1988 2004Included observations: 17VariableCoefficientStd. Errort-StatisticProb.C4.5654296.2086410.7353350.4790X1-0.5556870.235955-2.3550530.0403X20.0154920.0027235.6900120.0002X30.0201770.0305780.6598630.5242X40.1436280.0352784.0713200.0022X5-0.0101730.027910-0.3645020.7231X60.0
