Table 8.
Comparison of the proposed scheme with recently published ML models to predict COVID-19 patients' mortality risk
| Algorithm | Input features dataset | Key performance indicators (KPI) | |||||
|---|---|---|---|---|---|---|---|
| MPCD | Recall/sensitivity | F2 score | Precision | AUC | Accuracy | ||
| DL | 48 Clinical data | 0.93 | 1 | 0.93 | 0.91 | 0.93 | 0.95 |
| RF | 0.88 | 0.95 | 0.89 | 0.85 | 0.89 | 0.93 | |
| SVM | 0.77 | 0.91 | 0.87 | 0.87 | 0.87 | 0.89 | |
| ANN | 0.78 | 0.89 | 0.9 | 0.88 | 0.90 | ||
| XGBoost | 0.81 | 0.93 | 0.94 | 0.90 | 0.90 | 0.91 | |
| LR | 0.76 | 0.90 | 0.87 | 0.88 | 0.86 | 0.89 | |
| SVM and KNN [21] | 11 Clinical data | – | – | – | – | – | 0.80 |
| ANN [24] | 42 Clinical data | – | – | – | – | – | 0.90 |
| ML [27] | 12 Clinical data | – | 0.90 | – | – | 0.866 | – |
| DNN [28] | 51 Clinical data | – | 0.8125 | – | – | 0.97 | 0.9598 |
| ML [29] | 20 Clinical data | – | – | – | – | 0.94 | – |
| Multivariate Analysis [23] (Cox proportional regression) | 4 Clinical data | – | 0.95 | – | – | 0.91 | – |
| ML [34] | 3 Clinical data | – | – | – | – | 0.91 | – |
| Multivariate Regression model [35] | 7 Clinical data | – | – | – | – | 0.74 | – |
| CNN and Deep Transfer Learning [30] | RGB X-ray images | – | 0.9762 | – | – | – | 0.8810 |
| CNN and Deep Transfer Learning [31] | X-ray & CT-Scan images | – | 0.94 | – | 0.95 | – | 0.95 |
| Deep CNN-LSTM [54] | X-ray Images | – | 0.993 | – | – | 0.999 | 0.994 |
| CNN- Ensemble of Machine Learning [57] | X-ray Images | – | 0.978 | – | 1 | – | 0.989 |
| CNN-RNN [53] | X-ray Images | – | 0.999 | – | 0.999 | 0.999 | 0.999 |
| KNN [84] | Clinical data | 1.00 | 0.93 | 0.942 | 0.922 | 0.9374 | |