Thanks Dennis,
I don't think it was a problem of not feeding in a function for rollapply(),
because I was using mean() and my co.var() function in the FUN argument.
The key part seems to be the transformation that zoo() does to the matrix. If I
do the same transformation to my original matrix, the rollapply() function now
works fine.
Thanks again
David
Date: Fri, 15 Oct 2010 03:00:27 -0700
Subject: Re: [R] using apply function and storing output
From: djmu...@gmail.com
To: dasol...@hotmail.com
CC: r-help@r-project.org
Hi:
You need to give a function for rollapply() to apply :)
Here's my toy example:
d <- as.data.frame(matrix(rpois(30, 5), nrow = 10))
library(zoo)
d1 <- zoo(d) # uses row numbers as index
# rolling means of 3 in each subseries (columns)
> rollmean(d1, 3)
V1 V2 V3
2 3.000000 5.333333 4.333333
3 3.333333 4.333333 4.000000
4 2.666667 3.666667 3.333333
5 4.333333 3.666667 3.333333
6 5.000000 5.000000 5.000000
7 4.666667 4.666667 4.000000
8 5.666667 3.666667 4.000000
9 5.000000 3.000000 2.666667
# create new function
> cv <- function(x) sd(x)/mean(x)
# use new function in rollapply():
> rollapply(d1, 3, FUN = cv)
V1 V2 V3
2 0.3333333 0.1082532 0.5329387
3 0.1732051 0.4803845 0.6614378
4 0.2165064 0.5677271 0.4582576
5 0.7418193 0.5677271 0.4582576
6 0.6000000 0.3464102 0.4000000
7 0.7525467 0.4948717 0.6614378
8 0.8882158 0.5677271 0.6614378
9 1.0583005 0.3333333 0.2165064
This is a simplistic example, but it should get you started.
HTH,
Dennis
On Fri, Oct 15, 2010 at 2:00 AM, David A. <dasol...@hotmail.com> wrote:
Hi Dennis and Dimitris,
thanks for your answers. I am trying the rollmean() function and also the
rollapply() function because I also want to calculate CV for the values.
For this I created a co.var() function. I am having problems using them.
>co.var<-function(x)(
+sd(x)/mean(x)
+)
> dim(mydata)
[1] 1710 244
>xxx<-rollmean(mydata,3,by=3)
works fine and creates a vector which I will transform into a matrix. I still
have to find out how to store the output in my previously created 570x244 0's
matrix in an ordered way.
but, since the examples in the help page says it´s the same, I tried
> xxx<-rollapply(mydata,3,mean,by=3)
Error in UseMethod("rollapply") :
no applicable method for 'rollapply' applied to an object of class
"c('matrix', 'integer', 'numeric')"
and, with my created function...
> xxx<-rollapply(ord_raw_filt.df,3,FUN=co.var,by=3)
Error in UseMethod("rollapply") :
no applicable method for 'rollapply' applied to an object of class
"c('matrix', 'integer', 'numeric')"
Can you help me with the error?
Dave
Date: Fri, 15 Oct 2010 00:45:08 -0700
Subject: Re: [R] using apply function and storing output
From: djmu...@gmail.com
To: dasol...@hotmail.com
CC: r-help@r-project.org
Hi:
Look into the rollmean() function in package zoo.
HTH,
Dennis
On Fri, Oct 15, 2010 at 12:34 AM, David A. <dasol...@hotmail.com> wrote:
Hi list,
I have a 1710x244 matrix of numerical values and I would like to calculate the
mean of every group of three consecutive values per column to obtain a new
matrix of 570x244. I could get it done using a for loop but how can I do that
using apply functions?
In addition to this, do I have to initizalize a 570x244 matrix with 0's to
store the calculated values or can the output matrix be generated while
calculating the mean values?
Cheers,
Dave
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
[[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.
