Table 3.

Performance evaluation of different ML models using mrmr feature selection technique for NCS.

	Features	Accuracy (%)	Sensitivity (%)	Specificity (%)	F1-score (%)	Error rate	MCC	Kappa	AUC
EC	Top 10	93.25 ± 0.95	91.69 ± 1.03	98.44 ± 0.61	91.77 ± 1.02	0.07 ± 0.01	0.90	0.82	1.00
RF	Top 10	93.06 ± 0.63	91.61 ± 0.69	98.92 ± 0.59	91.52 ± 0.76	0.07 ± 0.01	0.89	0.81	1.00
DT	Top 10	91.34 ± 1.60	89.86 ± 1.88	99.37 ± 0.46	89.46 ± 1.99	0.09 ± 0.02	0.87	0.77	0.98
KNN	Top 10	79.47 ± 0.94	75.71 ± 0.89	91.95 ± 1.05	75.89 ± 1.01	0.21 ± 0.01	0.69	0.45	0.91
SVM	Top 8	75.98 ± 1.59	69.29 ± 1.92	75.18 ± 2.06	72.54 ± 1.84	0.24 ± 0.02	0.64	0.36	0.96
NB	Top 10	73.90 ± 2.02	72.35 ± 2.16	95.31 ± 1.01	72.43 ± 2.02	0.26 ± 0.02	0.64	0.30	0.95
LR	Top 9	71.76 ± 1.89	69.45 ± 1.85	93.42 ± 1.22	69.19 ± 1.82	0.28 ± 0.02	0.60	0.25	0.95
DAC	Top 9	70.73 ± 2.44	68.66 ± 2.43	94.11 ± 1.24	68.52 ± 2.24	0.29 ± 0.02	0.59	0.22	0.94