初学eview,看到有相关的文章,我找到了原始数据,也但是怎么做都不和文章上的结果不一致,希望高手帮帮忙:
1、方程为:
lr=A1lr(-1)+ε
h=B0+B1ε^2+B2σ^2
2、我的步骤:
(1)利用genr lr=log(r)
(2)在equation中选择arch,然后再均值方程中输入,lr lr(-1)
3、我的结果
Dependent Variable: LR
Method: ML - ARCH (Marquardt) - Normal distribution
Date: 11/25/07 Time: 22:33
Sample (adjusted): 2002M02 2006M12
Included observations: 59 after adjustments
Convergence achieved after 13 iterations
Variance backcast: ON
GARCH = C(2) + C(3)*RESID(-1)^2 + C(4)*GARCH(-1)
Coefficient Std. Error z-Statistic Prob.
LR(-1) 0.999704 0.000413 2423.355 0.0000
Variance Equation
C 6.99E-06 2.32E-06 3.005888 0.0026
RESID(-1)^2 -0.132133 0.100251 -1.318028 0.1875
GARCH(-1) 1.081914 0.082763 13.07241 0.0000
R-squared 0.897757 Mean dependent var 4.554731
Adjusted R-squared 0.892180 S.D. dependent var 0.039579
S.E. of regression 0.012996 Akaike info criterion -5.885319
Sum squared resid 0.009289 Schwarz criterion -5.744469
Log likelihood 177.6169 Durbin-Watson stat 1.697308
4、文章的结果:
参数
系数值 0.999576 1.74-06 -0.093563 1.077624
-统计量 2498.614 1.039688 -3.027965 27.12347
相伴概率 0.0000 0.2985 0.0025 0.0000
调整的r2 0.890889 AIC值-5.850944 SC值-5.710094
[此贴子已经被作者于2007-11-25 22:47:12编辑过]