[R] Predict VAR

[David Winsemius]

On May 25, 2010, at 4:49 PM, David Winsemius wrote:

>
> On May 25, 2010, at 4:14 PM, Luis Felipe Parra wrote:
>
>> Hello, I am using the predict function for VAR in r obtaining the  
>> following
>> object for the predictions with the following command
>
> Hard to tell what that means since you have not specified what  
> package was used or the class of the object on which predict was  
> applied.
>>
>> PronFac <- predict(VARFactores,n.ahead=1)
>>
>>> PronFactores$fcst
>
> The problem is that you have here shown a different object than what  
> you created with predict.
>
>> $PC1
>>            fcst      lower  upper       CI
>> PC1.fcst 2.284497 -0.8033048 5.3723 3.087802
>> $PC2
>>             fcst     lower    upper       CI
>> PC2.fcst -0.938333 -4.346927 2.470261 3.408594
>> $PC3
>>             fcst     lower    upper       CI
>> PC3.fcst -1.035569 -4.282719 2.211582 3.247151
>> $PC4
>>              fcst     lower    upper       CI
>> PC4.fcst -0.7063035 -4.027811 2.615204 3.321507
>> $PC5
>>             fcst     lower    upper       CI
>> PC5.fcst 0.3664593 -1.689041 2.421959 2.055500
>>
>>
>> I would like to take the fcst object from each of the list elements  
>> and
>> assign it to a vector, do you know how can I do this in an  
>> efficient way
>> without having to go trough all the list with a for?
>
> Loops are just as efficient as the apply methods.
>
>> At the moment I am
>> doing it the following way for(i in
>> 1:NumFac){PronFactores[i,]<-$fcst[[i]][1] } but I would like to know
>> if there is a commmand to acces to vectors of forecasts without the
>> confidence intervals. Does anybody know how to do this?
>
> Hard to tell for sure since you're only showing us output and not a  
> reproducible version. The str function is more informative than  
> printing to the console.

After looking at objects created in the help page for the predict  
function in the vars package, I suspect you want:

sapply(PronFac$fcst, "[", 1)  #not more efficient but more expressive  
perhaps.


>
>
>>
>> Thank you
>>
>> Felipe Parra
>
>
> David Winsemius, MD
> West Hartford, CT
>
> ______________________________________________
> R-help at 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.

David Winsemius, MD
West Hartford, CT