Table 1. Performance measures for the testing set data (n = 423).

For all measures, higher values represent better model predictions. Balanced accuracy considers the uneven balance of classes (‘temperate’ and ‘virulent’) and was calculated with ‘adjusted = True’ such that random guessing would have a score of 0 and a perfect model would have a score of 1. ‘MCC’ stands for Matthew’s correlation coefficient.

	BACPHLIP	Mavrich	PHACTS
Accuracy	0.983	0.955	0.79
Balanced accuracy	0.97	0.917	0.528
MCC	0.967	0.911	0.586
F1-score	0.985	0.939	0.837