OpenTox Virtual Conference 2021 Session 2

We extend previously defined model validation strategies for binary mixtures to the more complex case of general, N-ary mixtures. Additionally, we propose a method, related to the so-called X-randomization, of establishing a baseline model performance for each mixture dataset to be in used in model selection comparisons. This baseline is intended to account for the statistical dependence generically present in mixture datasets. We contend that without such a baseline, estimates of model performance can be dramatically overconfident.