Uncertainty quantification of cuffless blood pressure estimation based on parameterized model evidential ensemble learning

• DOI: 10.1016/j.bspc.2024.106104
• Publication Date: 2024-02-21T12:29:37Z
• Authors: Zhan Shen; Tapabrata Chakrab...; Wenyan Wang; Shuaiting Yao; Zhizhong Fu; Yifan Chen; Xiaorong Ding
• Keywords: Uncertainty Quantification; Ensemble forecasting; Ensemble Learning

Cuffless blood pressure (BP) measurement has the potential to break through way detect and prevent hypertension, but it is still challenging in meeting clinical performance requirements. The limited accuracy of current cuffless BP mainly attributed epistemic (model) aleatoric (data) uncertainties estimation methods. However, few previous studies have considered this problem. In study, we propose a parameterized model evidential ensemble learning (PEEL) framework with aim reduce uncertainty (so as improve performance) quantify uncertainty. PEEL consists two stages: original estimations models first stage, and, neural network estimate distribution final second stage. Experiments on 96 subjects MIMIC III dataset show that error for systolic diastolic 3.74 mmHg 2.22 mmHg, respectively. estimation. Furthermore, estimated can be used confidence interval assist diagnosing hypertension support decisions.