Table.

Hyperparameters for the automated RGC classifier

Parameter	Algorithm level	Optimized value	Optimization range
Number of features	elastic net	54	5 to 100
Number of folds	elastic net	6	2 to 10
Alpha	elastic net	.326	0.0 to 1.0
Number of lambda values	elastic net	6	5 to 50
Number of repetitions	decision node	5	5 to 20
Minimum size	decision node	13	5 to 100
Maximum depth	decision tree	6	2 to 8
Minimum tree count	adaboost forest	76	20 to 100
Maximum tree count	adaboost forest	95	25 to 100
Stopping criterion	adaboost forest	1.44% improvement over last 44 trees	1% to 50% improvement over last 10 to 50 trees
Ensemble size	ECOC	96	32 to 100
Probability of ensemble membership	ECOC	22.2% in positive class, 38.2% in negative class, 60.4% null	10% to 90% in positive/negative class