Skip to main content
. 2023 Jun 1;13:8874. doi: 10.1038/s41598-023-35649-9

Figure 4.

Figure 4

(a) Random Forest, (b) XGBoost. Feature importances of the models, cancer specific survival. The SHAP values show the most important features for the Random Forest and XGBoost models, allowing for analysis and validation, based on medical knowledge, of the algorithms’ training. Both presented similar columns among the top ten, with only the order varying, probably due to differences between the two algorithms.