0

I'm trying to calculate a matrix of interpersonal influence based on the probability of a tie between two actors and the susceptibility to influence (Friedkin, 2001). For doing so, you start by a 13*13 matrix of direct ties between 0 and 1. The conversion of the sociomatrix to an influence matrix takes places as follows:

Wij=AiCij/(∑kCik)

with Ai = measure of susceptibility and Cik = matrix of the probability of a tie - calculated by an ERGM. What I don't understand, and it might be an easy question for mathematicians, is the denominator. Through what exactly are we dividing by? And is there an easy way to write a function in R for this calculation?

Thanks for clarifying.

Best, Mathias

Mathias De Roeck
  • 11
  • 3

1 Answers1

0

if Cij is the matrix of tie probabilities, isn't Cik just the sum of the tie probabilities excluding j? basically, the row sum but iterated over each column excluding that column. so conceptually something like

Cik <- matrix(NA,n,n) #where n <- 13 or whatever
for(j in 1:n){
   tmp <- Cij
   tmp[,j] <- 0
   Cik[,j] <- rowSums(tmp)
}

there is surely a more elegant solution though

tvg
  • 388
  • 2
  • 13