|
AI- based algorithm |
Only model was proposed, no implementation |
[76]
|
|
CNN |
Accuracy (82.9%), specificity (80.5%), sensitivity (84%) |
[20]
|
|
CNN |
Accuracy (97.8%) |
[30]
|
|
CNN |
F1-score (0.89) |
[43]
|
|
CNN |
Sensitivity (100%), specificity (100%), accuracy (100%), F1- score (100%) |
[71]
|
| Disease detection |
CNN |
Accuracy (98%), recall (96%), specificity (100%) |
[31]
|
|
CNN, SVM |
Accuracy (98.27%), sensitivity (98.93%), specificity (97.60%) F-1 score (98.28%), precision (97.63%), Matthews correlation coefficient (96.54%) |
[32]
|
|
CNN,SVM |
Accuracy(95.38%), FPR(95.52%), F1- score(91.41%), kappa (90.76%) |
[77]
|
|
GAN Network |
Accuracy (99.9%) |
[37]
|
|
Decision trees, random forests, and support vector machines |
Accuracy(80%) |
[46]
|
|
CNN and grad cam |
Accuracy (90.8 %), AUC(0.949) |
[38]
|
|
Random forest, Lasso-elastic-net regularized generalized linear (GLMnet), ANN |
AUC(0.94) |
[39]
|
|
CNN |
Accuracy(97.36%) of classification accuracy, sensitivity (97.65%),precision (99.28%) |
[40]
|
| Epidemic forecasting |
Modified auto-encoder for modeling time Series |
The estimated average errors of 6, 7, 8, 9 and 10-step forecasting were 1.64%, 2.27%, 2.14%, 2.08%, and 0.73% respectively |
[19]
|
|
Epidemiological model and ML-based AI model |
The two models’ predictions and actual data were plotted in a graph and there was a fit between the actual and predicted data |
[50]
|
|
Machine Learning technique of Levenberg-Marquardt |
The coefficient of determination,  >0.5, which is higher for the proposed model for most of the countries |
[51]
|
| Sustainable development |
Regression analysis and Group method of Data Handling |
Accuracy (85.7%) |
[52]
|
|
CNN and grad cam |
AUC (0.989), sensitivity (98.2%) |
[41]
|
| Diseases diagnosis |
CNN |
AUC (0.96) |
[42]
|
|
XGBoost machine learning algorithm |
Death prediction accuracy (100%), survival prediction accuracy (90%) |
[44]
|
|
CNN |
Accuracy (0.8982) of Bayesian CNN for prediction, predictive entropy (0.99) as a measure of uncertainty |
[47]
|
|
Neural Network, random forest, decision tree (CRT) |
Sensitivity(88.0%,), specificity(92.7%), PPV(68.8%), NPV(97.7%), accuracy(92.0%), AUC (0.90) |
[48]
|
|
CNN |
Atazanavir shows inhibitory potency with Kd of 94.94 nM against the SARS-CoV-2 3C-like proteinase, Remdesivir (113.13 nM), Efavirenz (199.17 nM), Ritonavir (204.05 nM) and Dolutegravir (336.91 nM) |
[49]
|
