(taxometrics) A classification method based on the numerical analysis of the variation of a large number of characters in a group of organisms. It is assumed that a classification will be more predictive the more characters on which it is based. It is also assumed that, to begin with, each character is of equal weight (see weighting
) although some characters may later be weighted. Initially, a matrix of data is compiled of *operational taxonomic units
(OTUs) against characters so that for every OTU the state of each of perhaps 50 or more characters is recorded. This matrix can be subjected to a variety of mathematical analyses, which provide a measure of the similarity or dissimilarity between all the OTUs. The end product is usually one or more dendrograms.