Blood Pressure Estimation Using Photoplethysmography Only: Comparison between Different Machine Learning Approaches

Figure 5

(a) Simplified flow diagrams of MLR in which coefficients and the random error were updated in each iteration to converge the least square error function. (b) Flow diagrams of SVM regression. Epsilon and slack variables surrounding the hyperplane contributed to make the dual objective formula with the Lagrangian function to solve the equation for BP estimators. (c) Simplified flow diagrams of regression tree algorithms. Different algorithm combinations were used to derive least square function, prune, and split tree in to branch nodes. Each node (small black colour-filled circles) contains an estimation result.

(a)

(b)

(c)