A bagging approach for improved predictive accuracy of intradialytic hypotension during hemodialysis treatment

Chien-Liang Liu, Min-Hsuan Lee, Shan-Ni Hsueh, Chia-Chen Chung, Chun-Ju Lin, Po-Han Chang, An-Chun Luo, Hsuan-Chi Weng, Yu-Hsien Lee, Ming-Ji Dai, Min-Juei Tsai

Abstract: Highlights•Enhanced prediction of intradialytic hypotension improves pre-emptive treatment in hemodialysis.•UnderXGBoost method addresses extreme dataset imbalance in medical prediction models.•Validation shows that UnderXGBoost outperforms alternatives in predicting hypotension events.•Sensitivity analysis confirms the robustness of UnderXGBoost for clinical applications.•The proposed method offers a new solution to various medical problems with imbalanced data.