Novel Crow Swarm Optimization Algorithm and Selection Approach for Optimal Deep Learning COVID-19 Diagnostic Model

Table 8

The first dataset diagnostic performance outcomes of various pretrained models.

No	Classifier	AUC	CAR	F1-score	Precision	Recall	FPR	PPV	NPV	MSE
1	ResNet50	90.89	89.10	87.23	88.63	87.74	88.93	88.75	90.33	0.054
2	DarkNet	82.22	84.23	83.45	82.27	82.34	84.16	83.21	82.44	0.075
3	GoogleNet	84.76	86.76	84.91	85.20	85.10	85.72	83.33	84.36	0.059
4	MobileNetV2	83.12	85.33	84.19	84.37	83.49	84.87	85.10	83.22	0.047
5	Xception	84.83	81.35	80.12	79.99	80.43	80.84	80.32	80.98	0.061
6	VGG19	84.22	84.97	84.77	82.46	82.86	83.33	83.47	81.23	0.046
7	VGG16	91.34	89.96	88.75	88.15	88.95	88.97	89.14	87.99	0.038
8	InceptionV3	72.21	75.34	73.81	73.47	73.83	74.87	72.34	74.84	0.069
9	ResNet34	90.42	88.21	86.48	82.97	87.22	87.53	87.39	76.27	0.048
10	CNNs	85.29	83.65	83.10	82.55	81.76	82.40	82.77	82.74	0.050
11	DNN	81.96	83.50	82.37	83.12	82.12	83.12	81.94	82.84	0.083
12	SAE	83.10	82.90	81.87	80.43	81.27	81.84	81.36	81.38	0.078
13	InceptionResNetV2	88.38	88.35	87.43	86.76	86.42	87.53	88.05	87.46	0.069
14	LSTM	82.44	83.24	82.33	82.58	83.09	82.77	82.55	79.24	0.099
15	NASNet-Large	81.11	82.35	80.48	80.21	80.68	81.31	81.68	82.16	0.062