CODEWITHSUNDEEP
rmatrix
BEN
BEN

Reputation: 77

Calculation on a list of matrix

I have a list of matrix, e.g. list(M1,M2,M3,M4)

I would like to get a matrix that add all the matrix M1+M2+M3+M4 and divide each row by rowSums(M1)+rowSums(M2)+rowSums(M3)+rowSums(M4).

Here an example list, 

lst <- list(M1 = matrix(c(1,4,2,5), 2, 2), M2 = matrix(c(2,6,3,7), 2, 2))
lst
#> $M1
#>      [,1] [,2]
#> [1,]    1    2
#> [2,]    4    5
#> 
#> $M2
#>      [,1] [,2]
#> [1,]    2    3
#> [2,]    6    7

where my desired outcome would be,

(1+2)/((2+1)+(2+3)) = 0.375      (2+3)/((2+1)+(2+3)) = 0.625
(4+6)/((6+7)+(4+5)) = 0.4545455  (5+7)/((6+7)+(4+5)) = 0.5454545

Upvotes: 1

Views: 68

Answers (2)

jogo
jogo

Reputation: 12559

You can use a 3-dimensional array and apply()

M1 <- matrix(c(1,4,2,5), 2)
M2 <- matrix(c(2,6,3,7), 2)
L <- list(M1, M2)
M <- array(unlist(L), dim=c(2,2,2)) # or: M <- array(c(M1, M2), dim=c(2,2,2))

apply(M, 1:2, sum) / apply(M, 1, sum)
# > apply(M, 1:2, sum) / apply(M, 1, sum)
#           [,1]      [,2]
# [1,] 0.3750000 0.6250000
# [2,] 0.4545455 0.5454545

and in rowSums notation (thx to user20650): 

rowSums(M, dims=2) / rowSums(M)

Upvotes: 3

akrun
akrun

Reputation: 887048

We can try

Reduce(`+`, lst)/ sum(sapply(lst, rowSums))

Or 

Reduce(`+`, lst)/ sum(unlist(lst))

Update

Based on the update by the OP (using the OP's dataset)

Reduce(`+`, lst)/rowSums(sapply(lst, rowSums))
#          [,1]      [,2]
#[1,] 0.3750000 0.6250000
#[2,] 0.4545455 0.5454545

data

lst <- list(matrix(1:16, 4, 4), matrix(1:16, 4, 4))

Upvotes: 5

Related Questions