Table 4.

5-fold cross-validation of optimized models.

Fold	BERT			XLNet
	AUC	Precision	Recall	AUC	Precision	Recall
1	0.7452	0.4567	0.8126	0.7438	0.4592	0.8027
2	0.7366	0.5244	0.6783	0.7570	0.4740	0.8159
3	0.7703	0.4711	0.8640	0.7692	0.4873	0.8292
4	0.7432	0.5350	0.6849	0.7454	0.4496	0.8292
5	0.7689	0.5047	0.7993	0.7557	0.4815	0.7977
Mean	0.7528	0.4984	0.7678	0.7542	0.4703	0.8149

Performance metrics in the validation set, after 50 epochs.

The best mean score for each metric (AUC, precision, recall) is in bold.