A gradient boosting tree model for multi-department venous thromboembolism risk assessment with imbalanced data

Handong Ma, Zhecheng Dong, Mingcheng Chen, Wenbo Sheng, Yao Li, Weinan Zhang, Shaodian Zhang, Yong Yu

Abstract: Highlights•This study proposes a new perspective of VTE risk assessment for multiple departments as an MTL problem.•Two novel variants are proposed based on the state-of-the-art tree-based MTL model to handle data imbalance.•The proposed algorithms can boost the predictive performance effectively for both the overall and single-department tasks.