Table 2.
Prediction (classification) results for presepsin groups (positive, negative) based on different artificial intelligence classifier algorithms without missing data.
| Dataset (n = 173) | Algorithm | Precision | Recall | F1 Score | ROC AUC | Accuracy (Cross Validation, k = 5) |
|---|---|---|---|---|---|---|
|
routine laboratory parameters ∼ presepsin Classification |
k-NN | 0.77 | 0.77 | 0.77 | 0.81 | 0.80 |
| Logistic Regression | 0.93 | 0.92 | 0.93 | 0.98 | 0.91 | |
| Naive Bayes Classifier | 0.93 | 0.92 | 0.93 | 0.95 | 0.92 | |
| Random Forest | 0.95 | 0.94 | 0.94 | 0.97 | 0.91 | |
| XGBoost | 0.95 | 0.94 | 0.94 | 0.97 | 0.94 |