Table 5. . The misclassification rates, numbers of true positives, false positives, false negatives, true negatives, sensitivity, specificity and area under the curve of the models.

Method	Variables	Misclassification rate (%)	TP (n)	FP (n)	FN (n)	TN (n)	Sensitivity	Specificity	AUC
Best subset	right_distTop_value, conj_varAspect_value, conj_varY_value, conj_varYlef_value	22.4	7	56	1	191	0.875	0.773	0.827
Best subset	right_distTop_value, conj_varX_value	13.3	6	32	2	215	0.75	0.87	0.85
LASSO	right_distTop_value, conj_boxscore_value	18.0	6	45	2	202	0.75	0.818	0.841
Random forest	25 variables	23.1	6	55	2	192	0.75	0.777	–

AUC: Area under the curve; FN: False negative; FP: False positive; LASSO: Least absolute shrinkage and selection operator; TN: True negative; TP: True positive.