Table 2.

Classification metrics Precision, Recall, F1-score and Specificity for different classes.

Metric	Algorithm	KNN	ANN	SVM	DT	RF	AB
	Class
Precision	1	0.50	0.76	0.58	0.78	0.71	0.69
	2	0.49	0.53	0.58	0.54	0.54	0.53
	3	0.65	0.74	0.57	0.78	0.76	0.76
	4	0.40	0.50	0.58	0.60	0.49	0.53
	5	0.39	0.63	0.28	0.62	0.66	0.58
Recall	1	0.33	0.61	0.52	0.64	0.64	0.65
	2	0.61	0.73	0.27	0.68	0.58	0.54
	3	0.90	0.86	0.98	0.89	0.82	0.76
	4	0.32	0.37	0.39	0.39	0.47	0.55
	5	0.13	0.28	0.16	0.43	0.57	0.57
F1-score	1	0.40	0.68	0.55	0.70	0.67	0.67
	2	0.54	0.61	0.37	0.61	0.56	0.53
	3	0.76	0.80	0.72	0.83	0.79	0.76
	4	0.35	0.42	0.47	0.48	0.48	0.53
	5	0.20	0.39	0.21	0.50	0.61	0.58
Specificity	1	0.91	0.95	0.90	0.95	0.93	0.92
	2	0.80	0.79	0.94	0.82	0.84	0.87
	3	0.77	0.85	0.64	0.88	0.88	0.87
	4	0.96	0.97	0.98	0.98	0.96	0.95
	5	0.97	0.97	0.93	0.96	0.95	0.94

Classes are constituted as 1- Cobalt, 2- Caesium, 3- Iridium, 4- Uranium and 5- Thorium for the PGAA database used in this study.

Best metric values are indicated in bold.