Table 7.

Metrics for the training and testing (external) cohorts for the LR^a-TF-IDF^b model.

Training Data Size	AUCc score	Sensitivity	Specificity	Precision	F1-score
10,000	0.81	0.66	0.80	0.40	0.50
100,000	0.83	0.75	0.74	0.37	0.50
1,000,000	0.84	0.71	0.80	0.42	0.53

^aLR: logistic regression.

^bTF-IDF: term frequency–inverse document frequency.

^cAUC: area under the receiver operating characteristic curve.