[R-SIG-Finance] Fwd: fGarch
Spencer Graves
spencer.graves at pdf.com
Wed Jan 30 04:58:10 CET 2008
Your model "~garch(1,1)" specifies a constant + heteroscedastic noise.
The slot you printed 'fit at fitted' is the estimated (constant) mean of
the series. To see the estimated noise standard deviation, try the
following:
> fit at sigma.t
[1] 0.4942740 0.3945277 0.3415809 0.3210639 0.1287955 0.2227873 0.2152462
[8] 0.2816325 0.1532941 0.1202254 0.1291714 0.1460944 0.1200281 0.1303487
[15] 0.1341825 0.2970904 0.2980062 0.2529998 0.2168916 0.1197274 0.1221059
[22] 0.1997035 0.1257199 0.2609890 0.3220154 0.2870766 0.1438314 0.1280727
[29] 0.1270258 0.1597462 0.2861451 0.2407515 0.2998393 0.2143946 0.3378366
[36] 0.7663652 1.1671236 1.1711026 1.2397624 1.2168715
>
For a similar example, see section 3.5 of Tsay (2005) Analysis of
Financial Time Series (Wiley) and "scripts\ch03" in the FinTS package.
hope this helps.
Spencer Graves
babel at centrum.sk wrote:
> ______________________ Pôvodná správa: ________________________
>
>> Od: babel at centrum.sk
>> Komu: <r-sig-finance at stat.math.ethz.ch>
>> Datum: 29.01.2008 14:28
>> Předmět: [R-SIG-Finance] fGarch
>>
>>
> Hello.
> I have this problem. Why do I have all fitted values the same??
>
>> y
>>
> 43.097 43.041 43.019 42.769 42.533 42.542 42.466 42.817 42.734 42.770
> 42.637 42.710 42.669 42.782 42.993 42.994 42.944 42.902 42.714 42.746
> 42.881 42.760 42.489 42.422 42.460 42.641 42.675 42.678 42.827 42.981
> 42.930 42.996 42.899 43.037 43.478 43.882 43.886 43.955 43.932 43.998
>
>
>> library(fGarch)
>>
>
>
>> fit = garchFit(~garch(1, 1), data = y)
>>
>
>
>> show.fGARCH(fit)
>>
>
>
> Title:
> GARCH Modelling
>
> Call:
> garchFit(formula = ~garch(1, 1), data = y)
>
> Mean and Variance Equation:
> ~arma(0, 0) + ~garch(1, 1)
>
> Conditional Distribution:
> dnorm
>
> Coefficient(s):
> mu omega alpha1 beta1
> 42.30131209 0.00394317 0.99198425 0.05514354
>
> Error Analysis:
> Estimate Std. Error t value Pr(>|t|)
> mu 42.301312 0.008606 4915.050 <2e-16 ***
> omega 0.003943 0.001626 2.426 0.0153 *
> alpha1 0.991984 0.061981 16.005 <2e-16 ***
> beta1 0.055144 0.052464 1.051 0.2932
> ---
> Signif. codes: 0 `***´ 0.001 `**´ 0.01 `*´ 0.05 `.´ 0.1 ` ´ 1
>
> Log Likelihood:
> 3405.273 normalized: 1.706904
>
>
>
>
>> fit at fitted
>>
> 42.30131 42.30131 42.30131 42.30131 42.30131 42.30131 42.30131 42.30131
> 42.30131 42.30131 42.30131 42.30131 42.30131 42.30131 42.30131 42.30131
>
>
>> predict(fit, n.ahead = 10)
>>
> meanForecast meanError standardDeviation
> 1 42.30131 2.581365 8.109771
> 2 42.30131 2.581365 8.298906
> 3 42.30131 2.581365 8.492442
> 4 42.30131 2.581365 8.690480
> 5 42.30131 2.581365 8.893126
> 6 42.30131 2.581365 9.100487
> 7 42.30131 2.581365 9.312673
> 8 42.30131 2.581365 9.529796
> 9 42.30131 2.581365 9.751972
> 10 42.30131 2.581365 9.979319
>
> I want to count RMSE and choose which Garch model, is better, but I am not
> able to make a garch model.Thank you
>
> _______________________________________________
> R-SIG-Finance at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only.
> -- If you want to post, subscribe first.
>
> _______________________________________________
> R-SIG-Finance at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only.
> -- If you want to post, subscribe first.
More information about the R-SIG-Finance
mailing list