Table 5:
Summary of algorithms performance
| Accuracy [95%CI] |
Sensitivity | Specificity | NPV | PPV | P [Acc>NIR] |
AUC | |
|---|---|---|---|---|---|---|---|
| Testing data | |||||||
| Logistic regression | 67% [52%-81%] | 55% | 82% | 66% | 69% | 0.009 | 0.67 |
| Decision tree | 86% [71%-97%] | 77% | 84% | 72% | 80% | <0.001 | 0.79 |
| Random forest | 89% [78%-99%] | 85% | 95% | 79% | 82% | <0.001 | 0.83 |
| Training data | |||||||
| Logistic regression | 74% [65%-83%] | 65% | 85% | 74% | 75% | 0.003 | 0.84 |
| Decision tree | 88% [83%-95%] | 82% | 92% | 81% | 85% | <0.001 | 0.92 |
| Random forest | 100% [97%-100%] | 100% | 100% | 100% | 100% | <0.001 | 1.00 |
Neutrophil to lymphocyte ratio (NLR), Mean Platelet Volume (MPV), International Normalized Ratio (INR), Total Bilirubin (TB).