Table 6 Comparison between the proposed method and existing methods.
From: Conformal prediction quantifies wearable cuffless blood pressure with certainty
Sample size | Training set (Model Input; Model Output) | Model | BP | Error metrics (mmHg) | UQ | CIs | ||
|---|---|---|---|---|---|---|---|---|
MAD | ME ± SD | |||||||
Hae et al.31 | 1129 | Baseline clinical characteristics and 24-h ABPM; Follow-up mean BP | CatBoost | SBP | 8.30 | 8.40 ± 7.00 | – | – |
DBP | 5.30 | 5.30 ± 4.30 | – | – | ||||
Liu et al.32 | 1125 | ECG, PPG and PPW signals; ABPM | HGCTNet | SBP | 6.10 | − 0.40 ± 8.60 | – | – |
DBP | 5.20 | − 0.40 ± 7.00 | – | – | ||||
Cisnal et al.33 | 500 | PPG features; ABPM | GB | SBP | 11.35 | − 0.33 ± 14.54 | – | – |
DBP | 7.85 | − 0.22 ± 10.10 | – | – | ||||
This work | 483 | ECG and PPG features; ABPM | GBRT | SBP | 14.27 | − 0.19 ± 17.95 | \(\checkmark\) | \(\checkmark\) |
DBP | 9.83 | 0.57 ± 12.28 | \(\checkmark\) | \(\checkmark\) | ||||
