Fig. 6 | Scientific Reports

Fig. 6

From: SMOTE-augmented machine learning model predicts recurrent and metastatic breast cancer from microbiome analysis

Fig. 6

Bar plots of the top 10 most important features based on importance scores derived from feature importance, permutation analysis, and SHAP. (a–c) Feature importance for tree-based classifiers: (a) random forest, (b) gradient boosting, and (c) AdaBoost. (d) SHAP values for the non-tree-based classifier, nearest neighbors. (e) Feature importance using permutation analysis for Naïve Bayes. Bars represent the mean permutation importance scores, and error bars indicate the standard deviation. SHAP: Shapley additive explanation.

Back to article page