Fig. 6
Performance evaluation of the XGBoost model. (A) Log loss versus epoch curves. These curves illustrate the model’s performance on the training and test sets as the number of epochs increases. The convergence and stabilization of log loss for both sets suggest that the model is neither underfitting nor overfitting and demonstrates good generalization. (B) Precision-Recall curve. This curve evaluates the trade-off between precision (the proportion of true positive predictions among positive predictions) and recall (the proportion of true positives identified out of all actual positives). A higher area under the Precision-Recall curve indicates strong model performance, particularly for imbalanced datasets. (C) Learning curve. This plot shows how model accuracy changes with increasing training data. The decreasing gap between training and cross-validation accuracy as sample size increases indicates that the model generalizes well to unseen data. (D) SHAP values plot. This visualization highlights the contribution of individual features to the model’s predictions. Positive SHAP values increase the likelihood of a positive prediction, while negative values decrease it, providing interpretability to the model’s decisions.
