Fig. 3

Screening important genes associated with the prognosis of AML patients based on SVM-RFE and XGBoost models. (A) indicates that the SVM-RFE algorithm identified 22 important genes. The SVM-RFE algorithm filtered 26 genes with prognostic value to determine the best combination of feature genes. Finally, 22 genes (maximum accuracy = 0.9797) were identified as the optimal feature genes. (B) and (C) indicate that the XGBoost algorithm identified 19 important genes. (B) Importance scores of the top 19 important genes and corresponding variables screened by XGBoost. X-axis indicates the importance score which is the relative number of a variable that is used to distribute the data, Y-axis indicates the top 19 weighted variables (C) The ROC curve of the XGBoost model, The AUC (area under the ROC curve) value is 0.964, which indicates that the predictive performance of the XGBoost model is good