Calinski-Harabasz index — what is it?

medium

Answer

$\mathrm{CH}(k)\; = \;[\mathrm{Tr}(B_{k})\; / \;(k - 1)]\; / \;[\mathrm{Tr}(W_{k})\; / \;(n - k)]$
$Analogue to an F-statistic for clustering.$
$Fast, no ground truth needed.$
$Bias: tends to favor higher k.$
$Complements silhouette.$
$Combined with domain, one of the standard internal indices.$

Check yourself — multiple choice

Random
Ratio of between to within dispersion; higher = better; F-statistic analog for clustering; biased toward higher k
Same as silhouette
Not real

CH index: between/within dispersion ratio; higher better; F-analog.

#clustering#evaluation

214 interview questions in this topic.