TABLE 3.

Performance comparison of various ML techniques based on 12 morphological features for prediction of soybean charcoal rot disease.

Method	Accuracy	Sensitivity	Specificity	Precision	NPV	F1 score	MCC	AUC
LR-L1	95.58 ± 8.51	95.25 ± 10.27	95.92 ± 6.97	95.66 ± 7.59	95.61 ± 9.28	95.41 ± 8.97	91.21 ± 16.92	97.24 ± 5.84
LR-L2	94.96 ± 9.63	94.50 ± 12.51	95.42 ± 7.32	94.99 ± 8.32	95.15 ± 10.55	94.64 ± 10.54	90.03 ± 18.99	96.96 ± 6.40
MLP	94.50 ± 8.64	94.17 ± 12.04	94.83 ± 5.88	94.46 ± 6.57	94.90 ± 10.32	94.17 ± 9.41	89.18 ± 16.98	97.29 ± 5.45
RF	95.46 ± 9.33	95.08 ± 10.33	95.83 ± 8.41	95.64 ± 8.88	95.31 ± 9.73	95.35 ± 9.62	90.93 ± 18.63	97.12 ± 6.35
GBT	96.13 ± 7.64	95.92 ± 8.34	96.33 ± 6.97	96.22 ± 7.22	96.05 ± 8.04	96.06 ± 7.78	92.26 ± 15.27	98.00 ± 4.28
SVM	95.58 ± 7.73	95.67 ± 9.05	95.50 ± 6.53	95.34 ± 6.92	95.88 ± 8.49	95.48 ± 7.99	91.20 ± 15.41	97.46 ± 5.51