How do I create similarity matrices in r?

user_9671713 0
Viewed 327

I have a matrix with legislators on the rows, issues on the columns, and values indicating how legislators voted on a given issue (0 - absent, 1 - for, 2 - against). I need to create a similarity matrix (whether or not two legislators voted the same way if they were both present for the vote) for each issue. Is there a quick way of doing that in r without using nested for loops?

Here is how the matrix looks like:

    I1 I2 I3 I4
L1   1  1  1  2
L2   1  1  0  0
L3   2  2  2  2
L4   2  2  0  0

Here is what I would like to get for the first issue:

   L1 L2 L3 L4
L1     1  0  0
L2  1     0  0
L3  0  0     1
L4  0  0  1
similarity r matrix
1 Answers

You may try

func <- function(a, b) {ifelse(a==b & a*b != 0, 1, 0)}

lapply(dummy, function(x) {
  res <-outer(x, x, func)
  diag(res) <- NA
  colnames(res) = rownames(res) = c("L1", "L2", "L3", "L4")
  res
  })


$I1
   L1 L2 L3 L4
L1 NA  1  0  0
L2  1 NA  0  0
L3  0  0 NA  1
L4  0  0  1 NA

$I2
   L1 L2 L3 L4
L1 NA  1  0  0
L2  1 NA  0  0
L3  0  0 NA  1
L4  0  0  1 NA

$I3
   L1 L2 L3 L4
L1 NA  0  0  0
L2  0 NA  0  0
L3  0  0 NA  0
L4  0  0  0 NA

$I4
   L1 L2 L3 L4
L1 NA  0  1  0
L2  0 NA  0  0
L3  1  0 NA  0
L4  0  0  0 NA

data

dummy <- read.table(text = "    I1 I2 I3 I4
   1  1  1  2
   1  1  0  0
   2  2  2  2
   2  2  0  0", header = T)
rownames(dummy) <- c("L1", "L2", "L3", "L4")

   I1 I2 I3 I4
L1  1  1  1  2
L2  1  1  0  0
L3  2  2  2  2
L4  2  2  0  0
user_167291750
Related