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

Table 7

The first dataset diagnostic performance outcomes of various pretrained models.

No	Classifier	AUC	CAR	F1-score	Precision	Recall	FPR	PPV	NPV	MSE
1	ResNet50	90.78	91.46	90.49	89.73	88.94	90.92	90.22	90.17	0.039
2	DarkNet	80.92	85.13	85.11	83.29	80.14	82.19	83.71	81.95	0.069
3	GoogleNet	86.99	90.35	89.42	90.21	90.11	88.92	90.47	90.16	0.044
4	MobileNetV2	85.47	88.36	87.69	83.33	86.91	82.14	85.25	86.94	0.038
5	Xception	75.14	77.13	75.02	76.96	74.82	77.37	74.38	72.96	0.094
6	VGG19	80.32	84.38	84.29	77.43	93.56	82.14	83.64	81.79	0.081
7	VGG16	81.36	80.14	80.25	78.97	79.46	78.19	80.03	76.91	0.078
8	InceptionV3	63.40	65.98	64.81	66.47	64.91	62.99	65.11	63.73	0.055
9	ResNet34	90.54	90.71	80.48	79.28	80.10	89.79	90.21	89.94	0.042
10	CNNs	87.47	88.15	83.36	87.55	87.89	86.36	87.99	86.69	0.057
11	DNN	83.39	85.36	81.30	83.57	85.66	83.96	84.14	82.64	0.063
12	SAE	80.39	82.14	79.92	84.87	81.93	80.97	83.94	83.12	0.059
13	InceptionResNetV2	85.67	87.95	86.11	87.64	84.24	86.14	88.19	85.37	0.098
14	LSTM	88.25	90.54	88.36	89.25	88.97	86.91	90.11	87.67	0.096
15	NASNet-Large	79.36	80.11	78.98	77.91	78.16	75.87	79.47	82.96	0.079