TABLE 5.

Information that is useful to report when describing ML modeling efforts

Raw data processing	Contaminant removal, abundance/prevalence filtering, normalization, transformation. See references 15, 36.
Data splitting	Ratios of train, test, validation split; performance variation with different data splits; source of external test sets
Feature selection	Method of feature selection, when in process feature selection was applied, whether cross-folded
Model type	Type(s) of model(s) used and why, metrics used for model evaluation and optimization
Hyperparameters	Final hyperparameters selected for model, and methods used for hyperparameter optimization
Code	Open-source code and publicly available data are ideal but not always required