TABLE 2. Hyperparameters of Panic Attack Models.

Model	Parameter	Value
Random forest	n_estimators	100
	min_samples_split	2
	max_depth	1
Decision tree	min_samples_leaf	1
	min_samples_split	2
Linear discriminant analysis	solver	Lsqr
	shrinkage	auto
Regularized greedy forest	max_leaf	1000
	algorithm	RGF_Sib
	test_interval	100
AdaBoost	n_estimators	50
	learning_rate	1
Deep neural network	fully-connected layer 1	128
	fully-connected layer 2	256
	fully-connected layer 3	128
	fully-connected layer 4	64
	dropout gradient	0.2
	class_weight for negative	1.0
	class_weight for positive	1.5