Fig. 2 | Scientific Data

Fig. 2

From: A benchmark for machine-learning based non-invasive blood pressure estimation using photoplethysmogram

Fig. 2

MASE results for SBP (left) and DBP (right) with confidence intervals extracted with bootstrapping of (a) Sensors, (b) UCI, (c) BCG, and (d) PPGBP. The significance lines show the pairwise comparison of the best model against the same category models (inside) and across categories (outside). The significance (a = 0.001(***), 0.01(**), 0.05(*)) is measured by assessing if the 1-a confidence intervals of the models’ difference contain 0. ‘ns’ stands for not significant. Bonferroni’s correction is used for these multiple comparisons. We highlight with bold type the best result within each category.

Back to article page