. 2023 Jan 18;13:986. doi: 10.1038/s41598-022-27170-2

Table 5.

Performance statistics of $Δ$ SBP estimation using the models proposed. Results are given as median (IQR) computed across all folds of the LOSOCV. Entries in bold indicate the best performance for that metric.

Model Name	$ρ_{P}$	RMSE *	MAE *
Baseline reference	- (-)	10.46 (6.18)	7.40 (4.87)
LASSO+ ${OLS}_{PPG}$ ${RF}_{PPG}$	0.85 (0.44) 0.88 (0.26)	6.19 (4.05) 5.75 (4.73)	5.41 (2.41) 4.75 (3.66)
LASSO+ ${OLS}_{ECG}$ ${RF}_{ECG}$	0.64 (0.70) 0.69 (0.45)	7.48 (6.37) 6.79 (5.66)	6.54 (4.63) 5.28 (4.57)
${LASSO+OLS}_{PPG + ECG}$ ${RF}_{PPG + ECG}$	0.82 (0.40) 0.86 (0.23)	7.99 (5.74) 5.66 (4.76)	6.30 (4.62) 4.86 (4.29)
${LASSO+OLS}_{PAT}$ ${RF}_{PAT}$	0.83 (0.10) 0.80 (0.15)	6.09 (4.25) 6.39 (4.23)	4.91 (3.33) 5.21 (3.69)

Model Name

ρ_{P}

RMSE *

MAE *

Baseline reference

- (-)

10.46 (6.18)

7.40 (4.87)

LASSO+ ${OLS}_{PPG}$

${RF}_{PPG}$

0.85 (0.44)

0.88 (0.26)

6.19 (4.05)

5.75 (4.73)

5.41 (2.41)

4.75 (3.66)

LASSO+ ${OLS}_{ECG}$

${RF}_{ECG}$

0.64 (0.70)

0.69 (0.45)

7.48 (6.37)

6.79 (5.66)

6.54 (4.63)

5.28 (4.57)

${LASSO+OLS}_{PPG + ECG}$

${RF}_{PPG + ECG}$

0.82 (0.40)

0.86 (0.23)

7.99 (5.74)

5.66 (4.76)

6.30 (4.62)

4.86 (4.29)

${LASSO+OLS}_{PAT}$

${RF}_{PAT}$

0.83 (0.10)

0.80 (0.15)

6.09 (4.25)

6.39 (4.23)

4.91 (3.33)

5.21 (3.69)

* results given in units of mmHg.