Table 7.

Effect of the Firefly algorithm on accuracy and training time.

Dataset	Classifications	Metrics	Firefly Algorithm	SVM	LR	RF
First Dataset	Abnormal vs. Normal	Accuracy	Yes	0.967	0.969	0.923
			No	0.969	0.967	0.886
		Training time (s)	Yes	315.31	63.84	55.46
			No	452.92	81.4	68.38
	AMD vs. DME	Accuracy	Yes	0.909	0.926	0.851
			No	0.929	0.934	0.828
		Training time (s)	Yes	141.94	47.15	36.34
			No	307.82	86.83	52.67
	CNV vs. DRUSEN	Accuracy	Yes	0.965	0.977	0.934
			No	0.965	0.971	0.934
		Training time (s)	Yes	117.24	24.93	35.29
			No	308	64.17	61.06
Second Dataset	Abnormal vs. Normal	Accuracy	Yes	0.97	0.976	0.94
			No	0.973	0.976	0.953
		Training time (s)	Yes	11.71	7.17	9.33
			No	23.52	11.17	13.61
	AMD vs. DME	Accuracy	Yes	0.933	0.933	0.91
			No	0.936	0.927	0.916
		Training time (s)	Yes	9.36	5.2	4.23
			No	10.86	5.99	4.54