Table 3 Evaluation of the multi-class model with test data
From: Early prediction of patient discharge disposition in acute neurological care using machine learning
Precision | Recall | |||||||||
LR | SVM | KNN | XG | RF | LR | SVM | KNN | XG | RF | |
Home | 71% | 55% | 71% | 64% | 66% | 53% | 63% | 36% | 61% | 68% |
Death | 44% | 45% | 52% | 48% | 49% | 57% | 55% | 54% | 56% | 54% |
rehab | 23% | 27% | 42% | 27% | 36% | 30% | 27% | 38% | 31% | 46% |
Nursng | 36% | 38% | 40% | 36% | 33% | 44% | 42% | 32% | 43% | 22% |
F1 | AUC | |||||||||
LR | SVM | KNN | XG | RF | LR | SVM | KNN | XG | RF | |
Home | 58% | 60% | 44% | 69% | 70% | 85% | 78% | 81% | 85% | 86% |
Death | 47% | 27% | 44% | 44% | 47% | 74% | 63% | 71% | 73% | 77% |
rehab | 29% | 20% | 27% | 21% | 23% | 61% | 57% | 56% | 60% | 61% |
Nursng | 35% | 27% | 34% | 35% | 37% | 66% | 61% | 61% | 66% | 68% |