mexter {dprep} | R Documentation |
This function computes some external measures for cluster validation
mexter(true.labels, estimated.labels)
true.labels |
The observed class labels |
estimated.labels |
The estimaded class labels |
rand |
The rand measure |
jaccard |
The jaccard measure |
fandm |
The Fowlkes-Mallows measure |
hubert |
The Hubert measure |
Edgar Acuna
##---- Should be DIRECTLY executable !! ---- ##-- ==> Define data, use random, ##-- or do help(data=index) for the standard data sets. ## The function is currently defined as function (true.labels, estimated.labels) { ndat = length(true.labels) tab <- table(true.labels, estimated.labels) nidot <- apply(tab, 1, sum) ni2 = sum(nidot^2) ndotj <- apply(tab, 2, sum) nj2 = sum(ndotj^2) z = sum(tab^2) num <- z - ndat tempo <- (ni2 - ndat) * (nj2 - ndat) rdenom = ndat * (ndat - 1) return(list(rand = 1 + 2 * (z - 0.5 * (ni2 + nj2))/rdenom, jaccard = num/(ni2 + nj2 - z - ndat), fandm = num/sqrt(tempo), hubert = (rdenom * num - tempo)/sqrt(tempo * (rdenom - ni2 + ndat) * (rdenom - nj2 + ndat)))) }