This framework enhances risk assessment for chronic kidney disease by improving accuracy and interpretability, leveraging Bayesian optimization and explainable AI.
Key Points
The optimized framework achieved 92.4% accuracy and 97.7% ROC-AUC in chronic kidney disease risk assessment, significantly improving predictive performance.
Utilizing Bayesian optimization reduced computational time by 74%, facilitating efficient parameter tuning in the machine learning model.
The approach integrates explainable AI methods, including SHAP, providing transparent insights into model decisions for chronic kidney disease predictions.
Comprehensive evaluation maintains equitable performance across diverse patient populations, addressing biases commonly seen in risk stratification.