Table 5.

Performance comparison of three XGBoost^a models in detecting the subjective sense of moderate-intensive marijuana intoxication class.

MLb model	MIc precision	MI recall	MI F1-score	MI AUCd
XGBoost-MobiFit	0.89	0.76	0.82	0.99
XGBoost-Mobile	0.64	0.61	0.62	0.96
XGBoost-Fitbit	0.65	0.63	0.64	0.98

^aXGBoost: eXtreme Gradient Boosting.

^bML: machine learning

^cMI: moderate-intensive intoxication.

^dAUC: area under the curve.