Skip to main content
Fig. 2 | BMC Pulmonary Medicine

Fig. 2

From: A clinical data-driven machine learning approach for predicting the effectiveness of piperacillin-tazobactam in treating lower respiratory tract infections

Fig. 2

Feature selection curves for each model. (a) Cross-validated AUC scores versus the number of selected features for different models. The optimal number of features and corresponding AUC scores are annotated for each model. (b) SHAP values showing the impact of each feature on model predictions. Features are ranked by their absolute SHAP values, with blue indicating lower values and red indicating higher values. ALB and CRP showed the strongest impact on model predictions

Back to article page

BMC Pulmonary Medicine

ISSN: 1471-2466

Contact us