Table 5.

Metrics of the best model achieved for each TECA subscale both in the validation and test sets.

Subscale	Model	Features (n)			Validation set					Test set
		Eye-tracking	Behavioral	Total	Accuracy	Kappa	AUC	TPR	TNR	Accuracy	Kappa	AUC	TPR	TNR
TECA EU	kNN	12	7	19	0.69	0.3	0.67	0.81	0.51	0.58	0	0.44	0.75	0.25
TECA EJ	Random forest	4	10	14	0.76	0.49	0.81	0.87	0.65	0.75	0.47	0.83	0.86	0.6
TECA ES	Random forest	16	1	17	0.81	0.58	0.85	0.88	0.71	0.67	0.31	0.66	0.71	0.6
TECA PT	Random forest	14	5	19	0.82	0.63	0.87	0.83	0.83	0.83	0.67	0.83	0.83	0.83

The number of variables used by each model is divided according to the source (i.e., eye-tracking or behavioral data). The values shown per metric in the validation set are the mean values of the cross-validation iterations. TPR, true positive rate; TNR, true negative rate.