Hi,
I'm using "garchFit" function from packages fGarch, to fit some nose.
The weird thing is that in some cases the estimate for the alpha1
coefficient is equal to 1, which can't be otherwise the model is not stable.
Even more alpha1 = 1 makes impossible to simulate the model.
My model is an ARMA(1,5)/GARCH(2,0), could be a probelm coused by
overfitting? All my parameters seems to be significant...
Can anyone help me to undestand way this is happening?
Below there is the output from the function.
Thanks in advance , regards,
Filippo.
This is the output from the function:
Title:
GARCH Modelling
Call:
garchFit(formula = ~arma(1, 5) + garch(2, 0), data = Text.noise,
trace = F)
Mean and Variance Equation:
data ~ arma(1, 5) + garch(2, 0)
<environment: 0xca9e848>
[data = Text.noise]
Conditional Distribution:
norm
Coefficient(s):
mu ar1 ma1 ma2 ma3 ma4
-7.9290e-15 -4.1513e-01 -1.0000e+00 -6.4137e-01 5.0686e-01 2.0932e-01
ma5 omega alpha1 alpha2
-6.1404e-02 9.2491e-04 1.0000e+00 1.0000e+00
Std. Errors:
based on Hessian
Error Analysis:
Estimate Std. Error t value Pr(>|t|)
mu -7.929e-15 2.185e-05 0.000 1
ar1 -4.151e-01 2.091e-02 -19.851 < 2e-16 ***
ma1 -1.000e+00 1.996e-02 -50.088 < 2e-16 ***
ma2 -6.414e-01 3.833e-02 -16.732 < 2e-16 ***
ma3 5.069e-01 2.807e-02 18.056 < 2e-16 ***
ma4 2.093e-01 1.457e-02 14.369 < 2e-16 ***
ma5 -6.140e-02 1.345e-02 -4.567 4.96e-06 ***
omega 9.249e-04 1.093e-04 8.460 < 2e-16 ***
alpha1 1.000e+00 8.741e-02 11.441 < 2e-16 ***
alpha2 1.000e+00 7.845e-02 12.747 < 2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Log Likelihood:
1359.009 normalized: 1.014943
Description:
Wed Oct 23 17:15:30 2013 by user:
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
