[R] problem with predict from nnet package

[Mauricio Romero]

Hi,

 

I’m having mayor issues with predict from the nnet package.

I’m training a neural network for forecasting. I trained the network with
info from 1995 to 2009 and I want to forecast month by month 2010.(the
network forecasts one month at a  time).

 

Since I have to do iterative forecasting, im using predict several times
including, including the new forecast each time, but for some reason predict
doesn’t forecast after the first itereation.

 

Here is my code:

 

pronostico4=function(x,rezagosx,muestra, reps, tam, maxi,vo,net){

library(timeSeries)

library(nnet)

library(dummies)

x=as.timeSeries(x)

X=NULL

             for (k in 1:length(rezagosx)) {

             X=cbind(X, lag(diff(log(x)),rezagosx[k]))

             }

             

              pred=predict(net, X,type="raw")

              print(sum(is.na(pred)))

              pred[muestra]=diff(log(x))[muestra]

              pred=exp(diffinv(pred[is.na(pred)==0],xi=log(vo)) )

              muestra2=muestra

              muestra2[max(which(muestra2==1))+1]=TRUE

              print(length(pred))

              print(sum(muestra2))

              print(sum(is.na(X)))

              x[muestra2]=pred[1:sum(muestra2)]

              return(x)

              }

              

              

              muestra=(is.na(a[,"Precio.Promedio.Bolsa"])==0 &
a[,"año"]<2009)

              pred=a[,"Precio.Promedio.Bolsa"]

              pred[muestra==F]=NA

               pred=as.timeSeries(pred)

                    X=NULL

                for (k in 1:length(rezagosx)) {

             X=cbind(X, lag(diff(log(pred)),rezagosx[k]))

             }

              net=nnet2(diff(log(pred)),X,muestra,reps,tam,maxi)

 

 

              for (i in 1:12){

 
pred=pronostico4(pred,rezagosx,muestra,reps,tam,maxi,vo,net)

                  muestra[max(which(muestra==1))+1]=TRUE

                  } 

 

 

 

Mauricio Romero 

Quantil S.A.

Cel: 3112231150

 

"It is from the earth that we must find our substance; it is on the earth
that we must find solutions to the problems that promise to destroy all life
here"

 


        [[alternative HTML version deleted]]


______________________________________________
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.