[R-SIG-Finance] rugarch bug or my mistake?

alexios galanos @|ex|o@ @end|ng |rom 4d@c@pe@com
Wed Apr 1 20:21:01 CEST 2020


There is no discrepancy here. It is just the way the plot is generated 
to connect the last value in the data with the first value in the forecast.
Instead of plotting, you could just look at the data:

 > tail(f using model$modeldata$data,1)
[1] -0.016112
 > tail(f using model$modeldata$data,1)
[1] -0.016112
 > head(f using forecast$seriesFor)
     1970-04-14 16:00:00
T+1        -0.001220014
T+2         0.002245227
T+3         0.001834683
T+4         0.001518531
T+5         0.001275068
T+6         0.001087582

-Alexios

On 4/1/20 10:39 AM, travisgleith . wrote:
> This is my first post to this mailing list so I apologize if I am out of
> order here.
> 
> I tried to ask this question on an online forum but got shot down
> <https://stats.stackexchange.com/questions/457924/is-this-a-bug-in-rugarch-or-is-the-forecast-supposed-to-conceal-the-first-forec?noredirect=1#comment849108_457924>
> 
> library(magrittr)
> library(rugarch)
> data = c(0.00448915878069781,-0.00593011429139564,-0.00521660375977179,0.00915792713276595,0.0281576105073823,0.00312671172731763,0.00354620561645788,-0.0116968104863564,0.00885293595922043,0.0107479700328324,-0.00286887922892909,-0.0157860645324253,-0.00034496291769634,0.00834156289619709,-0.0118426954076351,-0.0245429072803506,0.00920654078852251,-0.0194853526917377,0.0101437726804425,0.0121099494070556,0.0035482064346366,-0.00229786350435077,-0.00733982114587459,-0.00459640558705143,-0.012915903683943,-0.019073989784277,0.0126645393062962,0.00721493361382838,0.00999175394461038,0.0133412046238938,0.00648737743848443,-0.00562686238554141,-0.00330281416051201,0.00799100742663728,0.000663186933227922,0.0148432054223313,0.00196669935473182,-0.000615357170258429,0.00113636710367157,-0.0194719951587255,0.0169620174351603,-0.0131164201273478,-0.0135647068283389,-0.00438317321680465,-0.0147236772598623,0.00448073182498596,-0.00388037967044466,0.0258503616763401,0.00971753055430585,-0.0079
>   0819347457519,-0.00755558584805335,-0.0171629980145799,0.0107626388061985,-0.0105764253253602,-0.00728398157712776,-0.00098385324839817,0.00401832316660708,-0.00123931628811594,0.0117077717740899,0.017917887402996,0.010911835646347,-0.00827245645002914,-0.00906077086137638,-0.0205094556293795,-0.00737675894931045,-0.00690210266337801,0.020337194824323,0.00700174492576033,0.000228651273513847,-0.0143788420968017,0.013170014653836,-0.0041840803265823,-0.00998018021662783,-0.00773937960715987,0.0150502886036981,-0.00281352468848173,-0.0106468266964306,0.00603505797098974,0.00448080249830074,0.00828145475545461,0.000440945562519113,-0.0100964576962305,-0.00160882489466677,-0.000137141351361142,0.00954377280816113,0.00208528678239533,0.0209006426760515,0.00156626715014618,-0.00691347744367798,-0.00860724032017463,0.0110129635734199,0.00446238531960588,-0.0112374698383101,-0.0276009587172153,0.000631167643944636,0.000455412294070562,-0.000428413112601861,-0.00468012712974619,-0.0132383598
>   723957,0.0272178690714789,0.0346151554938807,0.00428132065168324,-0.0135193549155817,-0.0161119996079523)
> 
> m_g_spec = ugarchspec(
>    mean.model = list(
>      armaOrder = c(1,2)
>      , include.mean = T
>      , archm = F
>      , arfima = F
>    )
>    , distribution.model = "norm"
>    , variance.model = list(model = "sGARCH", garchOrder = c(1,1)))
> 
> m_g_fit = ugarchfit(m_g_spec, data, solver = "hybrid")
> f = ugarchforecast(m_g_fit)
> plot(f, which = 1)
> 
> f using model$modeldata$data %>% tail(1) %>% {abline(h = .)}
> f using forecast$seriesFor[1] %>% {abline(h = ., col = "red")}
> 
> The above code generates a chart.
> The black horizontal line shows the last value of the input data. The red
> line shows the first value of the forecast. Where did the last value in the
> blue series come from? The blue series is supposed to show actual data, but
> this last value is not in the input data. It seems to be the first forecast
> value but it is not given in the forecast object. Any idea what is going
> on?
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-SIG-Finance using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions should go.
> 


More information about the R-SIG-Finance mailing list