R/metrics.R

In Rwclust: Random Walk Clustering on Weighted Graphs

Defines functions hk_similarity l1_norm retrieve_similarity_function

retrieve_similarity_function <- function(similarity) {
  switch(
    similarity,
    "hk" = hk_similarity
  )
}

l1_norm <- function(x,y) {sum(abs(x - y))}

hk_similarity <- function(x, y, k) {
    exp(2*k - l1_norm(x, y)) - 1
}

Try the Rwclust package in your browser