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

Table 14

Performance analysis with and without image preprocessing.

Model	With preprocessing					Without preprocessing
	Accuracy	F1-score	MAE	RMSE	Specificity	Accuracy	F1-score	MAE	RMSE+	Specificity
SVM	93.39%	0.92	0.229	0.054	0.96	91.11%	0.91	0.298	0.089	0.94
Random Forest	95.19%	0.95	0.218	0.049	0.97	94.23%	0.94	0.227	0.054	0.96
Decision Tree	93.12%	0.92	0.262	0.063	0.96	92.02%	0.92	0.265	0.077	0.94
Naive Bayes	72.69%	0.72	0.795	0.396	0.87	69.63%	0.69	0.803	0.412	0.78
KNN	92.49%	0.92	0.226	0.051	0.97	90.15%	0.90	0.321	0.093	0.92
CNN	90.01%	0.90	0.314	0.087	0.95	89.45%	0.89	0.356	0.097	0.92
AlexNet	94.59%	0.94	0.206	0.061	0.97	93.78%	0.93	0.226	0.049	0.94
VGG-16	93.69%	0.93	0.227	0.051	0.96	91.82%	0.92	0.225	0.051	0.93
Resnet50	91.59%	0.91	0.272	0.077	0.96	91.18%	0.91	0.299	0.065	0.93
InceptionV3	89.78%	0.89	0.313	0.082	0.95	87.43%	0.86	0.391	0.099	0.94
CNN+SVM	91.12%	0.91	0.281	0.082	0.97	88.73%	0.89	0.366	0.096	0.93
CNN+Random Forest	92.49%	0.93	0.227	0.052	0.97	89.99%	0.90	0.325	0.088	0.93
CNN+Decision Tree	90.99%	0.91	0.271	0.078	0.96	88.31%	0.88	0.347	0.093	0.93
CNN+Naive Bayes	82.85%	0.83	0.456	0.123	0.91	79.56%	0.80	0.478	0.178	0.87
CNN+KNN	91.89%	0.92	0.253	0.061	0.95	89.56%	0.89	0.312	0.079	0.90
AlexNet+SVM	96.69%	0.97	0.217	0.043	0.98	95.12%	0.95	0.217	0.047	0.93
AlexNet+Random Forest	96.09%	0.96	0.225	0.049	0.98	95.11%	0.95	0.213	0.047	0.95
AlexNet+Decision Tree	93.09%	0.93	0.225	0.050	0.97	92.45%	0.92	0.223	0.053	0.92
AlexNet+Naive Byes	83.13%	0.83	0.421	0.099	0.91	80.55%	0.81	0.492	0.153	0.86
AlexNet+KNN	93.39%	0.93	0.220	0.055	0.94	90.91%	0.91	0.279	0.050	0.91
VGG-16+SVM	94.59%	0.95	0.205	0.061	0.97	93.69%	0.93	0.221	0.045	0.93
VGG-16+Random Forest	95.19%	0.95	0.200	0.054	0.97	93.34%	0.93	0.22	0.049	0.94
VGG-16+Decision Tree	93.39%	0.93	0.214	0.043	0.96	91.23%	0.91	0.277	0.080	0.93
VGG-16+Naive Bayes	84.68%	0.85	0.419	0.083	0.92	82.87%	0.83	0.455	0.122	0.88
VGG-16+KNN	93.09%	0.93	0.261	0.062	0.96	92.45%	0.92	0.227	0.053	0.92
Resnet50+SVM	93.69%	0.94	0.227	0.050	0.97	91.78%	0.93	0.220	0.048	0.94
Resnet50+Random Forest	94.29%	0.94	0.201	0.059	0.97	86.45%	0.86	0.399	0.087	0.93
Resnet50+Decision Tree	92.19%	0.91	0.278	0.079	0.96	89.10%	0.89	0.369	0.101	0.93
Resnet50+Naive Bayes	87.08%	0.87	0.389	0.100	0.94	85.18%	0.85	0.402	0.077	0.90
Resnet50+KNN	91.89%	0.92	0.249	0.058	0.96	88.99%	0.89	0.337	0.088	0.91
InceptionV3+SVM	92.79%	0.93	0.220	0.047	0.99	89.99%	0.90	0.319	0.091	0.92
InceptionV3+Random Forest	91.89%	0.92	0.236	0.059	0.96	87.91%	0.88	0.320	0.079	0.93
InceptionV3+Decision Tree	90.69%	0.91	0.266	0.070	0.95	88.45%	0.88	0.340	0.091	0.91
InceptionV3+Naive Bayes	86.18%	0.86	0.411	0.078	0.93	84.72%	0.85	0.416	0.081	0.91
InceptionV3+KNN	91.29%	0.91	0.288	0.075	0.95	90.11%	0.90	0.311	0.090	0.93