Prediction of COVID-19 with Computed Tomography Images using Hybrid Learning Techniques

Table 12

Confusion matrix and classification report for proposed work, Resnet50 for feature extraction, and machine learning models for classification.

Confusion matrix						Classification report
Models	Category	COVID-19	CAP	Normal	Total	Category	Precision	Recall	F1Score	Specificity
Resnet50+SVM	COVID-19	104	4	3	111	COVID-19	0.94	0.95	0.94	0.94
	CAP	3	104	4	111	CAP	0.94	0.93	0.93	0.93
	Normal	3	4	104	111	Normal	0.94	0.94	0.94	0.94
	Total	110	112	111	333	Average	0.94	0.94	0.94	0.94
Resnet50+Random Forest	COVID-19	105	3	3	111	COVID-19	0.95	0.95	0.95	0.97
	CAP	3	104	4	111	CAP	0.94	0.95	0.95	0.97
	Normal	3	3	105	111	Normal	0.94	0.94	0.95	0.97
	Total	111	110	111	333	Average	0.95	0.95	0.95	0.97
Resnet50+Decision Tree	COVID-19	102	4	5	111	COVID-19	0.92	0.93	0.93	0.96
	CAP	4	103	4	111	CAP	0.93	0.92	0.93	0.96
	Normal	4	5	102	111	Normal	0.92	0.92	0.92	0.96
	Total	110	11	111	333	Average	0.92	0.92	0.92	0.96
Resnet50+Naive Bayes	COVID-19	97	7	7	111	COVID-19	0.87	0.87	0.87	0.94
	CAP	7	96	8	111	CAP	0.88	0.86	0.87	0.94
	Normal	8	6	97	111	Normal	0.86	0.86	0.86	0.93
	Total	112	109	112	333	Average	0.87	0.86	0.86	0.94
Resnet50+KNN	COVID-19	102	5	4	111	COVID-19	0.92	0.92	0.92	0.95
	CAP	5	101	5	111	CAP	0.91	0.92	0.91	0.96
	Normal	4	4	103	111	Normal	0.93	0.92	0.93	0.96
	Total	111	110	112	333	Average	0.92	0.92	0.92	0.96