Figure 3From: Estimation of cardiac output and systemic vascular resistance using a multivariate regression model with features selected from the finger photoplethysmogram and routine cardiovascular measurementsPooled histogram of selected features and number of features used when estimating SVR using the best multivariate model. Top: Pooled HIST1 - relative frequency of a feature being selected to estimate SVR, bottom: Pooled HIST2 - relative frequency of number of features being used in the SVR estimation model. Two features were most commonly selected by the feature selection algorithm; feature 24 (log(LF/HF)) and feature 28 (log(PW)). Feature 3 (LF/HF) was sometimes selected.Back to article page