What is consensus clustering?

hard

Answer

$Run clustering many times (varying algorithm, k, subsample, init).$
$M_{\mathrm{ij}}\; = \;\mathrm{fraction}$
$Cluster M (usually hierarchical) \to consensus assignment.$
$Standard in bioinformatics for gene expression subtype discovery.$
$More robust than any single run; helps identify natural k when peaks in M are stable across k.$

Check yourself — multiple choice

Random
Repeat clustering, build co-clustering matrix $M_{\mathrm{ij}}$ , cluster M for consensus assignment; robust; standard in bioinformatics
Same as k-means
Not real

Consensus clustering: co-clustering matrix over many runs → consensus.

#clustering#evaluation

214 interview questions in this topic.