[R] Forescasting using predict() in an object of class arimax when there is an outlier IO in the model.

roman_elessar1 roman_lord2002 at hotmail.com
Wed Nov 16 00:40:17 CET 2011


Forescasting using predict() in an object of class arimax when there is an
outlier IO in the model.
Hi R users
I have a problem when a use the predict() method in an object of class
arimax ( These objects are the results of the implementation of the function
arimax() from the TSA library) . The object is a model of a time series in
which I identified an IO oulier at the element 33 of the serie (using the
function detectIO() from the package TSA in an object arimax that
corresponds a model ARIMA(1,1,0) that  not include the outlier) , so I
incorporated this in the model ARIMA(1,1,0):
mz3B<-arimax(z3B,io=c(33),order=c(1,1,0))
where the object z3B is a time series of class ts . When I want to use the
predict() method, there is a problem, and a warning message is printed:
> predict(mz3B,n.ahead=12)
Error in dim(data) <- dim : Attempting to specify an attribute in a NULL

So i dont know how to resolve this problem. Maybe a need to use another
function to make the forecast, but I don’t have any idea. Thanks in advance
for any responders.
The data I used is the next:
             Jan      Feb      Mar      Apr       May     Jun       Jul       
Aug    Sep      Oct
2008  36.54   23.03   24.97   23.80   42.62   31.92   22.43  33.59  28.30 
49.08
2009  45.59   38.38   37.85   31.93   30.52   30.79   33.62  38.25  27.35 
34.21
2010  38.18   39.70   40.17   33.85   36.98   41.44   44.97  47.91  74.31 
53.78
            Nov       Dec
2008  38.99   46.06
2009  37.66   41.96
2010  61.31   70.46


--
View this message in context: http://r.789695.n4.nabble.com/Forescasting-using-predict-in-an-object-of-class-arimax-when-there-is-an-outlier-IO-in-the-model-tp4074693p4074693.html
Sent from the R help mailing list archive at Nabble.com.



More information about the R-help mailing list