Fig. 3
From: Survival analysis using machine learning in transplantation: a practical introduction
RSF model confusion matrix. The confusion matrix output shows that the model has an accuracy of 89.67%, meaning it correctly classifies 89.67% of the instances. The sensitivity (recall) is high at 98.22%, indicating the model effectively identifies true positives. However, the specificity is low at 14.66%, meaning it struggles to correctly identify true negatives. The positive predictive value (precision) is 90.99%, showing that most positive predictions are correct, while the negative predictive value is 48.39%, indicating less reliability in negative predictions. The balanced accuracy, which averages sensitivity and specificity, is 56.44%, reflecting the model's performance on imbalanced data