Table 4.

F1 scores for each respective model and outcome in a 1-month and 3-month follow-up period after participants’ initial onboarding.

Outcome and model		F1 Score, %
		1-month follow-up	3-month follow-up
HIVa
	Support vector machine	70.5	81.6
	Random forest	73.4	82.6
	Naïve Bayes	69.6	82.1
	Logistic regression	68.5	81.4
Amphetamines
	Support vector machine	88.2	85.9
	Random forest	88.3	85.9
	Naïve Bayes	88.0	85.5
	Logistic regression	88.1	85.8
Methamphetamines
	Support vector machine	88.3	85.1
	Random forest	88.3	85.3
	Naïve Bayes	88.1	84.8
	Logistic regression	88.2	85.0
THCb
	Support vector machine	11.1	7.4
	Random forest	5.0	0.5
	Naïve Bayes	24.0	4.7
	Logistic regression	10.9	9.1

^aHIV: human immunodeficiency virus.

^bTHC: tetrahydrocannabinol.