Table 3.

Performance comparison of the exploratory models for low social interaction frequency and high levels of loneliness^a.

Exploration goal and model			AUCb		Accuracy		Precision		Specificity		F1-score
											Macroc		Microd
Low social interaction frequency
	GBMe	0.909		0.850		0.837		0.857		0.829		0.828
	LRf	0.850		0.780		0.837		0.732		0.763		0.766
	RF g	0.935		0.849		0.837		0.857		0.824		0.828
	XGBoosth	0.907		0.840		0.814		0.857		0.814		0.814
High levels of loneliness
	GBM	0.887		0.838		0.871		0.784		0.867		0.871
	LR	0.804		0.779		0.839		0.676		0.825		0.825
	RF	0.909		0.818		0.871		0.730		0.854		0.857
	XGBoost	0.858		0.798		0.839		0.730		0.838		0.839

^aThe italicized values indicate the machine learning models with the best performance within each category (ie, low social interaction frequency and high levels of loneliness).

^bAUC: area under the receiver operator characteristic curve.

^cAverage F₁-score for 10 folds.

^dCalculated as the sum of the confusion matrix of folds.

^eGBM: Gradient Boosting Machine.

^fLR: logistic regression.

^gRF: random forest.

^hXGBoost: Extreme Gradient Boosting.