Fig. 1From: The maximum entropy principle for compositional dataThe Compositional Maximum Entropy (CME) approach. a Through maximum entropy, CME infers the unknown generative model of the observed component abundances. b. \(h_i\) embodies the influence of each (i) component. Components with large \(h_i\) tend to have higher abundances than those with small \(h_i\). \(K_{ij}\) embodies the interaction between pairs of components. Pairs with \(K_{ij}>0\) tend to coexist, while pairs with \(K_{ij}<0\) tend to be mutually exclusiveBack to article page