R help - Mar 2012 - Sliding a Window in R

Dear all,
I am having a vector of around 300.000 elements and I Want to slide fast a
window from the first element until the last-Windowsize

what I have so far is the following for statement:

 for (i in 1:(length(data[,1]) - windowSize)) {
        out[i] <- mean(data[i:(i + windowSize - 1), ])
        elements[i]<-length(i:(i + windowSize - 1))
      }

but this of course takes ages to run, especially with small window sizes!.
Is it possible to speed up this in many cores in R? If yes how?

I would like to thank you in advance for your help

B.R
Alex

	[[alternative HTML version deleted]]

Just move it to C and you'll probably be ok. I believe runmean in
library(caTools) provides a very fast implementation.

Michael

On Mar 3, 2012, at 8:31 AM, Alaios <alaios at yahoo.com> wrote:
> Dear all,
> I am having a vector of around 300.000 elements and I Want to slide fast a
window from the first element until the last-Windowsize
> 
> what I have so far is the following for statement:
> 
> ?for (i in 1:(length(data[,1]) - windowSize)) {
> ??? ??? out[i] <- mean(data[i:(i + windowSize - 1), ])
> ??? ??? elements[i]<-length(i:(i + windowSize - 1))
> ??? ? }
> 
> but this of course takes ages to run, especially with small window sizes!.
> Is it possible to speed up this in many cores in R? If yes how?
> 
> I would like to thank you in advance for your help
> 
> B.R
> Alex
> 
>    [[alternative HTML version deleted]]
> 
> ______________________________________________
> 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.

On Sat, Mar 3, 2012 at 8:31 AM, Alaios <alaios at yahoo.com>
wrote:
> Dear all,
> I am having a vector of around 300.000 elements and I Want to slide fast a
window from the first element until the last-Windowsize
>
> what I have so far is the following for statement:
>
> ?for (i in 1:(length(data[,1]) - windowSize)) {
> ??? ??? out[i] <- mean(data[i:(i + windowSize - 1), ])
> ??? ??? elements[i]<-length(i:(i + windowSize - 1))
> ??? ? }
>
> but this of course takes ages to run, especially with small window sizes!.
> Is it possible to speed up this in many cores in R? If yes how?
>
> I would like to thank you in advance for your help
>
rollmean in the zoo package is quite fast.  There are also a variety
of other rolling routines there, principally rollapply which is very
flexible but, of course, is much slower.

Please provide complete self-contained reproducible code next time as
per last two lines of every message in r-help.

-- 
Statistics & Software Consulting
GKX Group, GKX Associates Inc.
tel: 1-877-GKX-GROUP
email: ggrothendieck at gmail.com

On 03-03-2012, at 14:31, Alaios wrote:
> Dear all,
> I am having a vector of around 300.000 elements and I Want to slide fast a
window from the first element until the last-Windowsize
> 
> what I have so far is the following for statement:
> 
>  for (i in 1:(length(data[,1]) - windowSize)) {
>         out[i] <- mean(data[i:(i + windowSize - 1), ])
>         elements[i]<-length(i:(i + windowSize - 1))
>       }
> 
> but this of course takes ages to run, especially with small window sizes!.
> Is it possible to speed up this in many cores in R? If yes how?
Try this

library(zoo)
out <- rowSums(rollmean(data,windowSize))/windowSize

I don't think you need elements  since all entries have the same size.

Berend

On 03-03-2012, at 14:55, Berend Hasselman wrote:
> 
> On 03-03-2012, at 14:31, Alaios wrote:
> 
>> Dear all,
>> I am having a vector of around 300.000 elements and I Want to slide
fast a window from the first element until the last-Windowsize
>> 
>> what I have so far is the following for statement:
>> 
>> for (i in 1:(length(data[,1]) - windowSize)) {
>>        out[i] <- mean(data[i:(i + windowSize - 1), ])
>>        elements[i]<-length(i:(i + windowSize - 1))
>>      }
>> 
>> but this of course takes ages to run, especially with small window
sizes!.
>> Is it possible to speed up this in many cores in R? If yes how?
> 
> Try this
> 
> library(zoo)
> out <- rowSums(rollmean(data,windowSize))/windowSize

Correction:

out <- rowSums(rollmean(data,windowSize))/ <number of columns in data>

Berend