Outer product from matrix by row and a vector

An embedded and charset-unspecified text was scrubbed...
Name: not available
URL: <https://stat.ethz.ch/pipermail/r-help/attachments/20120808/692f7bf1/attachment.pl>

Dear all,

I am facing a challenge when applying the outer product with a  
matrix by rows.

What I have is a rather big matrix, which I would like to convert  
into a different matrix by doing something like

outer(matrix_row, vector, function(x,y) x[1]+5*x[4]/y)
In order to get there, I tried to get the matrix into a list using  
the mefa package, as subsetting the original data appears too  
complicated. However, if I apply the following code I always get a  
dimensional error, as the unlist() function is applied to the whole  
list.

library(mefa)
a<-matrix(c(1:6),ncol=3)
b<-c(1,2,3,4)
a<-mat2list(a,MARGIN=1)

outer(a,b,function(x,y){unlist(a)[1]+unlist(a)[3]/b})
(You really should not be referring to a and b inside that function  
body. It's arguments are now named x and y.)

Can you tell us what the answer _should_ be for the first entry? And a  
step by step illustration of how you got there?

If this is not what it should be, then you need to read up on the  
expected result of the division of a scalar by a vector in R.

  apply(a, 1, function(x) { x[1] + x[3]/ b } )

          [,1] [,2]
[1,] 6.000000  8.0
[2,] 3.500000  5.0
[3,] 2.666667  4.0
[4,] 2.250000  3.5

Any idea how to do this
Do "this" is still somewhat nebulous.
more elegantly than a nested for-loop?

Many thanks,

Ingo

David Winsemius, MD
Alameda, CA, USA
Can you post what you want your answer to be for the a and b you have given already?
---------------------------------------------------------------------------
Jeff Newmiller                        The     .....       .....  Go Live...
DCN:<jdnewmil at dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live Go...
                                      Live:   OO#.. Dead: OO#..  Playing
Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
/Software/Embedded Controllers)               .OO#.       .OO#.  rocks...1k
--------------------------------------------------------------------------- 
Sent from my phone. Please excuse my brevity.

Dear all,

I am facing a challenge when applying the outer product with a matrix
by rows.

What I have is a rather big matrix, which I would like to convert into
a different matrix by doing something like

outer(matrix_row, vector, function(x,y) x[1]+5*x[4]/y)

In order to get there, I tried to get the matrix into a list using the
mefa package, as subsetting the original data appears too complicated.
However, if I apply the following code I always get a dimensional
error, as the unlist() function is applied to the whole list.

library(mefa)
a<-matrix(c(1:6),ncol=3)
b<-c(1,2,3,4)
a<-mat2list(a,MARGIN=1)

outer(a,b,function(x,y){unlist(a)[1]+unlist(a)[3]/b})

Any idea how to do this more elegantly than a nested for-loop?

Many thanks,

Ingo

[[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.
Hi Jeff, David, 

what I'm trying to do is speed the currently used nest for loop

a<-matrix(c(1:6),ncol=3)
b<-c(1,2,3,4)

result<-matrix(rep(0, times=dim(a)[1]*length(b)),nrow=dim(a)[1])
for(ii in 1:dim(a)[1]){
  for(jj in 1:length(b)){
    result[ii,jj]<-a[ii,1]+a[ii,3]/b[jj]
      }
}

giving the result

     [,1] [,2] [,3]
[1,]    1    3    5
[2,]    2    4    6

Thanks for taking a look. 

Cheers, 

Ingo
Can you post what you want your answer to be for the a and b you have given already?
---------------------------------------------------------------------------
Jeff Newmiller                        The     .....       .....  Go Live...
DCN:<jdnewmil at dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live Go...
                                      Live:   OO#.. Dead: OO#..  Playing
Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
/Software/Embedded Controllers)               .OO#.       .OO#.  rocks...1k
---------------------------------------------------------------------------
Sent from my phone. Please excuse my brevity.

Ingo Reinhold <ingor at kth.se> wrote:

>Dear all,
>
>I am facing a challenge when applying the outer product with a matrix
>by rows.
>
>What I have is a rather big matrix, which I would like to convert into
>a different matrix by doing something like
>
>outer(matrix_row, vector, function(x,y) x[1]+5*x[4]/y)
>
>In order to get there, I tried to get the matrix into a list using the
>mefa package, as subsetting the original data appears too complicated.
>However, if I apply the following code I always get a dimensional
>error, as the unlist() function is applied to the whole list.
>
>library(mefa)
>a<-matrix(c(1:6),ncol=3)
>b<-c(1,2,3,4)
>a<-mat2list(a,MARGIN=1)
>
>outer(a,b,function(x,y){unlist(a)[1]+unlist(a)[3]/b})
>
>
>Any idea how to do this more elegantly than a nested for-loop?
>
>Many thanks,
>
>Ingo
>
>       [[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.
Hi Jeff, David, 

what I'm trying to do is speed the currently used nest for loop

a<-matrix(c(1:6),ncol=3)
b<-c(1,2,3,4)

result<-matrix(rep(0, times=dim(a)[1]*length(b)),nrow=dim(a)[1])
for(ii in 1:dim(a)[1]){
  for(jj in 1:length(b)){
    result[ii,jj]<-a[ii,1]+a[ii,3]/b[jj]
      }
}

giving the result

     [,1] [,2] [,3]
[1,]    1    3    5
[2,]    2    4    6
Hi.

The printed matrix is "a". The above code yields on my computer

       [,1] [,2]     [,3] [,4]
  [1,]    6  3.5 2.666667 2.25
  [2,]    8  5.0 4.000000 3.50

Try the following

  out <- a[, 1] + a[, 3] %o% (1/b)

  max(abs(out - result))

  [1] 4.440892e-16

Hope this helps.

Petr Savicky.

On Wed, Aug 08, 2012 at 06:03:19AM +0000, Ingo Reinhold wrote:
Hi Jeff, David,

what I'm trying to do is speed the currently used nest for loop

a<-matrix(c(1:6),ncol=3)
b<-c(1,2,3,4)

result<-matrix(rep(0, times=dim(a)[1]*length(b)),nrow=dim(a)[1])
for(ii in 1:dim(a)[1]){
 for(jj in 1:length(b)){
   result[ii,jj]<-a[ii,1]+a[ii,3]/b[jj]
     }
}

giving the result

    [,1] [,2] [,3]
[1,]    1    3    5
[2,]    2    4    6
That is not what result equals (as Petr illustrates.)
Hi.

The printed matrix is "a". The above code yields on my computer

      [,1] [,2]     [,3] [,4]
 [1,]    6  3.5 2.666667 2.25
 [2,]    8  5.0 4.000000 3.50

Try the following

 out <- a[, 1] + a[, 3] %o% (1/b)

 max(abs(out - result))

 [1] 4.440892e-16

Hope this helps.
It's also the case that the printed result is just the transpose of  
what I earlier offered as a possible implementation of the OP's pseudo- 
code.:

 > identical( t( apply(a, 1, function(x) { x[1] + x[3]/ b } ) ), result)
[1] TRUE
David Winsemius, MD
Alameda, CA, USA