还剩5页未读，继续阅读
文本内容:
思考与练习答案简述题略
二、单项选择题1-5:BAABC6-10:DCBDA
三、多项选择题1-5:ABCD、AD、ACD、ACD、AC6-10:ABCD、ABCD、AB、BCD、ABC11-13:CD、AC、ABC
四、判断正误题1-5:错误、错误、正确、错误、错误6-10:正确、错误、错误、正确、错误11-13:正确、正确、错误
五、填空题
1.滞后现象
2.分布滞后模型；自回归模型
3.Bo；6i；Zf=o Pi
4.经验加权估计法；阿尔蒙法
5.权数的设置主观随意性较大
6.递减型
7.多项式次数
8.最小
9.库伊克模型自适应预期模型局部调整模型
10.一阶自回归模型
11.自回归
12.杜宾h检验LM检验
13.工具变量法广义差分法
六、计算操作题
1.1由DW值可知，不存在自相关性2由式
7.
3.18可知，y*=1-5,y*=
0.6740,则6=
0.
32602.经验加权法，可以选择若干类型滞后结构尝试，具体略阿尔蒙法:输入命令CROSS XY结果如下，可选择滞后期4Sample:19962018Included observations:23Correlations areasymptotically consistentapproximationsY,X+i laglead
0.
98900.
98900.
87030.
89090.
75160.
76550.
63170.
64910.
50570.
51830.
35110.
36110.
20310.
21830.
06930.0779-
0.0522-
0.0468-
0.1495-
0.1481-
0.2166-
0.2373-
0.2895-
0.3222-
0.3351-
0.3704Variable Coefficient Std.Error t-Statistic Prob.C-
644.
1974830.5026-
0.
7756720.4500PDL
010.
0055490.
0094830.
5850820.5672PDL02-
0.
0167200.002851-
5.
8652690.0000PDL
030.
0038280.
0047480.
8063060.4327R-squared
0.984546Mean dependent var
19910.76Adjusted R-squared
0.981455S.D.dependent var
15002.94S.E.of regression
2043.082Akaike infocriterion
18.26697Sum squaredresid62612790Schwarz criterion
18.46580Log likelihood-
169.5362Hannan-Quinn criter.
18.30062F-statistic
318.5430Durbin-Watson stat
1.312960ProbF-statistic
0.000000*Lag Distributionof XCoefficient Std.Error t-StatisticI
00.
054300.
010575.
1388910.
026100.
005474.
7687720.
005550.
009480.585083-
0.
007340.00562-
1.305614-
0.
012580.01167-
1.07809Sum ofLags
0.
066030.
0034719.0253下方系数T检验效果不好，可进一步调整，选择由二1,输入LS YC PDLX,4,1,见下Included observations:19after adjustmentsVariableCoefficient Std.Error t-Statistic Prob.C-
572.
0879816.5960-
0.
7005760.4936PDL
010.
0131740.
00068519.
221030.0000PDL02-
0.
0168690.002813-
5.
9960420.0000R-squared
0.983876Mean dependent var
19910.76Adjusted R-squared
0.981861S.D.dependent var
15002.94S.E.of regression
2020.621Akaike infocriterion
18.
2041418.35326Sum squaredresid65326550Schwarz criterion
18.22937Log likelihood-
169.9393Hannan-Quinn criter.
1.350427F-statistic
488.1641Durbin-Watson statProbF-statistic
0.000000Lag Distributionof Xi Coefficient Std.Error t-Statistic
00.
046910.
005209.
0162910.
030040.
0024212.
415120.
013170.
0006919.22103-
0.
003690.00330-
1.118364-
0.
020560.00610-
3.37211Sum ofLags
0.
065870.
0034319.2210拟合效果提升
3.Sample adjusted:2007Q22020Q1Included observations:52after adjustmentsVariableCoefficientStd.Error t-Statistic Prob.C
4697.
9122107.
1532.
2295060.0304X
0.
0844880.
0315472.
6781590.0100Y-
10.
4010720.
1727052.
3222930.0244R-squared
0.722804Mean dependentvar
26824.16Adjusted R-squared
0.711490S.D.dependentvar
9969.164S.E.of regression
5354.749Akaike infocriterion
20.06532Sum squaredresid
1.40E+09Schwarz criterion
20.17789Log likelihood-
518.6982Hannan-Quinn criter.
20.10847F-statistic
63.88517Durbin-Watson stat
1.996326ProbF-statistic
0.000000由于模型中含有被解释变量的滞后期作为解释变量，通过DW检验无法检验序列自相关性，但可以利用LM检验其自相关性，下图显示，模型存在自相关性Breusch-Godfrey SerialCorrelation LMTest:F-statistic
42.43247Prob.F2,
470.0000Obs*R-squared
33.46589Prob.Chi-Square
20.0000可以进一步调整，观察变量趋势图，存在季节波动，引入若干滞后期，效果均不理想,随后对变量X、Y分别进行季节调整，变量窗口proc按钮下选择菜单Census X12调整后的变量分别记为oY_SA、X_SA,调整后再建立自回归模型LS Y_SA CX_SA Y_SA-1如下图:D叩endent Variable:Y_SAMethod:Least SquaresDate:08/29/21Time:17:40Sample adjusted:2007Q22020Q1Included observations:52after adjustmentsVariableCoefficientStd.Error t-Statistic Prob.C
2310.
381792.
32132.
9159650.0053X_SA
0.
0635560.
0230882.
7527850.0083Y_SA-
10.
5994690.
1332794.
4978630.0000R-squared
0.963686Mean dependentvar
26814.87Adjusted R-squared
0.962204S.D.dependentvar
9205.404S.E.of regression
1789.638Akaike infocriterion
17.87338Sum squaredresid
1.57E+08Schwarz criterion
17.98595Log likelihood-
461.7078Hannan-Quinn criter.
17.91653F-statistic
650.1767Durbin-Watson stat
2.306981ProbF-statistic
0.000000再经LM检验，不存在自相关性。
个人认证
优秀文档
获得点赞 0
