Figure 1
From: Noninvasive prediction of Blood Lactate through a machine learning-based approach
Model estimates of blood lactate concentration compared with the measured value. (A–C) are low combined with moderate intensity CWR tests. (D–F) are those for high-intensity exercise. (G–I) is the model obtained when the data of three different intensities are combined. Scatter plot are demonstrated in (A,D,G). Agreement using Bland-Altman plot are shown in (B,C,E,F,H) and (B,E,H,I) show the difference between estimated and measured value. (C,F,I) are the percentage of estimation error. The dark solid horizontal lines in each Bland-Altman plot represent average bias whereas the dotted lines stand for average bias ± 1.96 standard deviation (95% upper and lower limit). The standard deviation in (B,E,H) are 0.52, 1.83 and 2.51 mmol/liter. The standard deviation in (C,F,I) are 0.1, 0.2 and 0.53.
