mescla.emma.rank.rule_of_one¶
- mescla.emma.rank.rule_of_one(eigenvalues)[source]¶
Number of eigenvalues >= 1 (Kaiser’s criterion; Joreskog et al., 1976).
On a correlation matrix the eigenvalues sum to the number of species
m, so “eigenvalue >= 1” is identical to “explains more than 1/m of the variance” – the form used by Tubau et al. (2014), who with 11 species required an eigenvector to explain more than 9.1%.Always returns at least 1: a rank-0 mixing model is not a model.