Table 4 Diagnostic performance of each model for SMPP in internal validation cohort
Model | AUC | Precision | F1 score | Sensitivity | Specificity | Negative | Kappa |
|---|---|---|---|---|---|---|---|
| Â | Â | prediction rate | Â | ||||
LightGBM | 95.00% | 91.67% | 93.62% | 95.65% | 94.59% | 97.22% | 0.8951 |
XGboost | 92.50% | 89.36% | 90.32% | 91.30% | 93.24% | 94.52% | 0.842 |
Logistic | 88.33% | 84.78% | 84.78% | 84.78% | 90.54% | 90.54% | 0.7532 |
RandomForest | 95.00% | 95.45% | 93.33% | 91.30% | 97.30% | 94.74% | 0.8934 |
KNN | 81.67% | 78.57% | 75.00% | 71.74% | 87.84% | 83.33% | 0.6057 |
SVM | 88.33% | 83.33% | 85.11% | 86.96% | 89.19% | 91.67% | 0.7552 |
Decision Tree | 92.50% | 87.76% | 90.53% | 93.48% | 91.89% | 95.77% | 0.8433 |
Naïve Bayes | 76.67% | 73.68% | 66.67% | 60.87% | 86.49% | 78.05% | 0.4897 |