Table 7 Responsiveness and reliability in predicting length of stay, discharge disposition, and mortality for the best performing three models when trained with situational variables only
| Â | Reliability (Accuracy) | Responsiveness (AUC) | ||||
|---|---|---|---|---|---|---|
Training | Testing | Validation | Training | Testing | Validation | |
LOS | ||||||
 LSVM | 62.07% | 61.78% | 62.04% | 0.587 | 0.586 | 0.584 |
 CHAID | 64.92% | 64.88% | 65.10% | 0.648 | 0.651 | 0.649 |
 Decision List | 84.45% | 84.61% | 84.26% | 0.559 | 0.556 | 0.561 |
Discharge | ||||||
 LSVM | 79.81% | 79.81% | 79.53% | 0.581 | 0.581 | 0.581 |
 CHAID | 79.81% | 79.81% | 79.53% | 0.585 | 0.582 | 0.583 |
 Decision List | 90.97% | 91.05% | 91% | 0.557 | 0.553 | 0.554 |
Mortality | ||||||
 XGT Boost Tree | 99.87% | 99.89% | 99.89% | 0.5 | 0.5 | 0.5 |
 LSVM | 99.87% | 99.89% | 99.89% | 0.639 | 0.494 | 0.514 |
 Decision List | 58.31% | 58.29% | 58.53% | 0.571 | 0.455 | 0.536 |