Continuous Blood Pressure with Photoplethysmography: Machine Learning for Photoplethysmography Signal Processing

I have performed a comparative analysis of different methods of machine learning and statistical methods of modeling blood pressure from PPG data. I determined that a random forest regression using scikit-learn performed the best for the data we collected. This model was able to predict mean arterial pressure (MAP) with a mean absolute error of 4.654, well within the range of being potentially medically useful.