Table 3.

Machine learning model performance for prediction of L. monocytogenes stress response categories.

	Stress Type *
Model	Acid		Cold		Salt		Desiccation
GBM	0.87 ${}^{abc}$	(0.83–0.89)	0.97 ${}^a$	(0.96–0.98)	0.89 ${}^a$	(0.87–0.90)	0.89 ${}^{ab}$	(0.86–0.90)
RF	0.87 ${}^{ab}$	(0.86–0.88)	0.97 ${}^a$	(0.95–0.98)	0.89 ${}^a$	(0.87–0.90)	0.91 ${}^a$	(0.88–0.92)
SVMR	0.89 ${}^c$	(0.88–0.89)	0.97 ${}^a$	(0.96–0.98)	0.83 ${}^b$	(0.81–0.84)	0.83 ${}^c$	(0.80–0.84)
SVML	0.85 ${}^a$	(0.84–0.87)	0.96 ${}^a$	(0.94–0.97)	0.85 ${}^b$	(0.83–0.86)	0.88 ${}^{ab}$	(0.86–0.90)
NN	0.72 ${}^d$	(0.68–0.78)	0.96 ${}^a$	(0.93–0.98)	0.63 ${}^c$	(0.57–0.68)	0.69 ${}^d$	(0.56–0.76)
LB	0.89 ${}^{bc}$	(0.88–0.90)	0.97 ${}^a$	(0.97–0.98)	0.85 ${}^{ab}$	(0.83–0.88)	0.86 ${}^{bc}$	(0.85–0.88)

* Mean (range); means within a column with similar lower case superscript letter are not significantly different; random forest (RF), support vector machine (radial (SVMR) and linear (SVML) kernels), gradient boosting (GBM), neural network (NN) and logit boost (LB) models.