Table 2. Performance comparison of unsupervised classifiers on internal and external radiological reports.

Average performance values are reported.

Classifier	Internal Testing Set (n = 3715)			External Testing Set (n = 456)
	Precision	Recall	f1-score	Precision	Recall	f1-score
fastText+GMM	0.42	0.65	0.51	0.88	0.84	0.84
ivis-2D+GMM	0.83	0.80	0.80	0.88	0.88	0.88
ivis-10D+GMM	0.82	0.80	0.81	0.91	0.91	0.92
CheXpert Labeller	0.81	0.81	0.81	0.93	0.93	0.92