R help - Jun 2003 - Questions for package ts prediction

Dear helpers,

I am trying to write a function to return prediction values using
package ts. I have written three different versions since I am not sure
what's wrong with my func2. func and func1 return the same results.But
func1 and func2 don't. In particular, the only difference between
"func1" and "func2" is the function variable name being y
and data,
respectively.  But running the last line of the following script will
give the message:

Error in ts(x): object is not a matrix.

I am confused. Also, could somebody kindly let me what's the answer if
any for the following sunspot example from the package help:

data(sunspot)
(sunspot.ar <- ar(sunspot.year)) 
# why not just sunspot.ar <- ar(sunspot.year) ?
predict(sunspot.ar, n.ahead=25)

Thanks in advance.

Zhu Wang
Statistical Science Department
Southern Methodist University

(214)768-2453
-- 
zhu wang <zhuw at mail.smu.edu>

# time series prediction

func<-function(data)
  {(esti<- ar(data))
   return(predict(object=esti,newdata=data,n.head=5))
 }
func1<-function(y)
  {(esti<- ar(y))
   return(predict(esti,n.head=5))
 }
func2<-function(data)
  {(esti<- ar(data))
   return(predict(esti,n.head=5))
 }

y<-arima.sim(model=list(ar=c(1.7,-0.8)),n=100)
func(y)
func1(y)
func2(y)

You have a scoping problem: the predict method needs to find the data:
please supply an explicit newdata argument.  Your first example works
because `y' happens to be globally visible and be the right object.

On 9 Jun 2003, zhu wang wrote:
> I am trying to write a function to return prediction values using
> package ts. I have written three different versions since I am not sure
> what's wrong with my func2. func and func1 return the same results.But
> func1 and func2 don't. In particular, the only difference between
> "func1" and "func2" is the function variable name being
y and data,
> respectively.  But running the last line of the following script will
> give the message:
> 
> Error in ts(x): object is not a matrix.
> 
> I am confused. Also, could somebody kindly let me what's the answer if
> any for the following sunspot example from the package help:
> 
> data(sunspot)
> (sunspot.ar <- ar(sunspot.year)) 
> # why not just sunspot.ar <- ar(sunspot.year) ?
Have you tried it?  Please do so, and you will learn the difference!
This idiom is widely used in the R help files.
> predict(sunspot.ar, n.ahead=25)
-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595