Table 2.
Comparison between the results of machine learning and classical statistics.
| Variable | Machine learning | Univariate LR* | Multivariate LR* |
|---|---|---|---|
| Donor age | ✓ | ✓ | ✓ |
| MMF therapy | ✓ | ✓ | ✓ |
| 3-month eGFR | ✓ | ✓ | ✓ |
| Acute rejection | ✓ | ✓ | |
| Acute kidney injury | ✓ | ✓ | |
| CMV infection | ✓ | ||
| Length of the 1st hospitalization | ✓ | ||
| Hypertension | ✓ | ||
| Transfusion | ✓ | ||
| Dialysis duration | ✓ | ||
| Readmissions 1st year | ✓ | ✓ | |
| Delayed graft function | ✓ | ✓ | |
| Azathioprine therapy | ✓ | ||
| HLA MM A | |||
| HLA MM DR | |||
| Dialysis modality | |||
| Proteinuria |
MMF mycophenolate mofetil, eGFR estimated glomerular filtration rate, CMV cytomegalovirus, HLA human leucocyte antigen, MM mismatch.
*Bivariate and multivariate logistic regression (SPSS 25).