Table 6.

Metrics for the training and testing (external) cohorts for the Bio-Clinical-BERT^a model.

Training data size	AUCb score	Sensitivity	Specificity	Precision	F1-score
10,000	0.82	0.76	0.74	0.36	0.49
100,000	0.84	0.74	0.77	0.39	0.51
1,000,000	0.85	0.39	0.96	0.67	0.50

^aBERT: Bidirectional Encoder Representations from Transformers.

^bAUC: area under the receiver operating characteristic curve.