Hamming distance measure for pvclust

Question

I am trying to create a Hamming distance measure for the pvclust clustering method. (There isn't one defined for this function.) I'm based on the example given for the cosine measure:

cosine <- function(x) {
x <- as.matrix(x)
y <- t(x) %*% x
res <- 1 - y / (sqrt(diag(y)) %*% t(sqrt(diag(y))))
res <- as.dist(res)
attr(res, "method") <- "cosine"
return(res)
}

I try to do it this way:

hamming <- function(x) {
x <- as.matrix(x)
y <- t(x) %*% x
res <- sum(y != y)
res <- as.dist(res)
attr(res, "method") <- "hamming"
return(res)
}

Unfortunately it doesn't work properly. Anyone have any postings, where is the error and how to fix it?

If I try to use this function for a matrix, I get the value 0. — Avter, Jun 14 '22 at 16:23

Mohamed Desouky · Accepted Answer · 2022-06-14T16:47:41.667

0

Try this

hamming <- function(x) {
x <- as.matrix(x)
y <- (1 - x) %*% t(x)
res <- y + t(y)
res <- as.dist(res)
attr(res, "method") <- "hamming"
return(res)
}

edited Jun 14 '22 at 16:47

answered Jun 14 '22 at 16:33

Mohamed Desouky

4,340
2
4
19

I am getting an error: Error in command 'x != y': mismatched arrays – Avter Jun 14 '22 at 16:38
what about this edit . – Mohamed Desouky Jun 14 '22 at 16:47
@Avter please provide a reproducible example of your data. It's very hard for someone to provide you an answer without one. – user438383 Jun 14 '22 at 16:48
Thank you Dohamed Desouky, everything is working fine. – Avter Jun 14 '22 at 17:17

Hamming distance measure for pvclust

1 Answers1