Table 7.
Comparison of proposed system and previous studies for features affecting the mortality
| Author(s) | Algorithm | Key features | Results | |
|---|---|---|---|---|
| Feature | Percentage | |||
| Yan, Zhang et al. [10] | XGBoost machine learning algorithm | Male | 58.7% |
Male, fever, cough, fatigue, dyspnoea, lactic dehydrogenase (LDH), lymphocyte and high-sensitivity C-reactive protein (hs-CRP) are the key features for differentiating between critical patients from the two classes |
| Fever | 49.9% | |||
| Cough | 13.9% | |||
| Fatigue | 3.7% | |||
| Dyspnoea | 2.1% | |||
| Shuai Zhang et al. [26] | Univariable Cox regression Model | Age, years | – | Age, male, fever, cough, weakness, severely ill, any and hypertension are the most important factors affecting the mortality |
| Male | 60% | |||
| Fever | 66.67% | |||
| Cough | 70% | |||
| Weakness | 53.33% | |||
| Severely ill | 96.67% | |||
| Any | 70% | |||
| Hypertension | 53.33% | |||
| Cox_COVID_19 prediction system | Cox regression method | Age | – | Age, fever, cough, pneumonia, muscle pain and throat pain are the most important factors affecting the mortality |
| Male | 61.69% | |||
| Fever | 46.56% | |||
| Pneumonia | 36.7% | |||
| Cough | 29.86% | |||
| Throat Pain | 8.3% | |||