Skip to content

sum

10 messages · Jerome Asselin, Barry Rowlingson, Sundar Dorai-Raj +6 more

Luis Silva
# 
Dear helpers

I have a list where each element is a matrix (the list is 
obtained with lapply). I want to sum those matrices. Is there a 
function to do that? The sum function sums all the elements...
--


http://adsl.sapo.pt
Jerome Asselin
# 
The only way I can see is to vectorize your list of matrices. Here's an 
example.

matlist <- lapply(1:10,function(i) matrix(rnorm(12),3,4))
summat <- matrix(sapply(matlist,I)%*%rep(1,10),3,4)

You can use the loop below to verify that the above solution is correct.

forsum <- matrix(0,3,4)
for(i in 1:10)
forsum <- forsum+matlist[[i]]

"summat" and "forsum" should be the same.

HTH,
Jerome
On April 23, 2003 10:57 am, Luis Silva wrote:
Barry Rowlingson
#
Luis Silva wrote:
Since your matrix elements must be all the same shape (rows x columns) 
for you to sum them, you could store them in an array (see ?array) 
rather than a list. Then all you need do is apply 'sum' to two 
dimensions of the array.

  Here's a one-liner that converts your list into an array (by unlisting 
it and then packing into an array with the right three dimensions) and 
then runs apply(...,c(1,2),sum) to get the answer you want:

  First some sample data:

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

[[2]]
           [,1]      [,2]      [,3]
[1,] 0.1666667 0.5000000 0.8333333
[2,] 0.3333333 0.6666667 1.0000000

[[3]]
      [,1] [,2] [,3]
[1,]  0.5  0.5  0.5
[2,]  0.5  0.5  0.5

[[4]]
      [,1] [,2] [,3]
[1,] 0.01 0.01 0.01
[2,] 0.01 0.01 0.01

 > 
apply(array(unlist(foo),c(dim(foo[[1]])[1],dim(foo[[1]])[2],length(foo))),c(1,2),sum)
          [,1]     [,2]     [,3]
[1,] 1.676667 4.010000 6.343333
[2,] 2.843333 5.176667 7.510000

I'm sure there's a better way.

Baz
Sundar Dorai-Raj
#
Luis Silva wrote:
How about:

R> r = list(matrix(1:4,2,2),matrix(5:8,2,2),matrix(9:12,2,2))
R> expr = paste("r[[",seq(along = r),"]]", collapse="+")
R> eval(parse(text = expr))
      [,1] [,2]
[1,]   15   21
[2,]   18   24


Regards,
Sundar
Spencer Graves
# 
Can you create a 3-d array rather than a list?  If yes, "apply" will work.

hth.  spencer graves
Jerome Asselin wrote:
Roger D. Peng
# 
How about something like:
[,1] [,2] [,3] [,4]
[1,]    3    7   11   15
[2,]    4    8   12   16
[3,]    5    9   13   17
[4,]    6   10   14   18
-roger
_______________________________
UCLA Department of Statistics
http://www.stat.ucla.edu/~rpeng
On Wed, 23 Apr 2003, Luis Silva wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Deepayan Sarkar
    

    

      
      #
    

  


  

      
On Wednesday 23 April 2003 12:57 pm, Luis Silva wrote:

            

      
      
      
    

        
If the list is not too long, perhaps

sumlist <- 
    function(x) 
        if (length(x) > 2) sumlist(x[1:2]) + 
             sumlist(x[-(1:2)]) else do.call("+", x) 



sumlist(<whatever your list is>)


-Deepayan

  
  


  


      

    

    

      
      

        


  

        

      John Fox
    

    

      
      #
    

  


  

      
Dear Luis,

        
On Wednesday 23 April 2003 12:57 pm, Luis Silva wrote:

        
> I have a list where each element is a matrix (the list is
 > obtained with lapply). I want to sum those matrices. Is there a
 > function to do that? The sum function sums all the elements...

Here's a recursive function that works with a list of two or more matrices:

     sumMatrices <- function(matrices){
         if (length(matrices) > 2) matrices[[1]] + Recall(matrices[-1])
         else matrices[[1]] + matrices[[2]]
         }

Regards,
  John



-----------------------------------------------------
John Fox
Department of Sociology
McMaster University
Hamilton, Ontario, Canada L8S 4M4
email: jfox at mcmaster.ca
phone: 905-525-9140x23604
web: www.socsci.mcmaster.ca/jfox

  
  


  


      

    

    

      
      

        


  

        

      Peter Dalgaard
    

    

      
      #
    

  


  

      
Barry Rowlingson <B.Rowlingson at lancaster.ac.uk> writes:

            

      
      
      
    

        
Didn't someone do an abind() function at some point? (Generalizing
cbind/rbind)

            

      
      
      
    

        
l <- foo
apply(array(unlist(l),c(dim(l[[1]]),length(l))),c(1,2),sum)

is a bit shorter.

Even shorter, although I'm not sure better in all senses of the word,
same basic idea:

matrix(apply(sapply(l,c),1,sum),dim(l[[1]]))

However, aren't we ignoring the obvious?:

s<-0;(for(a in l)s<-s+a)