Classification of Photoplethysmographic Signals using Support Vector Machines for Vascular Risk Assessment

Rohan Baid, Niranjana Krupa, and Muhammad A.M. Ali


Cardiovascular diseases, PPG, ARX, linear parametric model, support vector machine


Cardiovascular diseases have registered a high rate of morbidity and mortality in the world, therefore the assessment of cardiovascular risk in human beings is of prime importance. In this paper Photoplethysmographic (PPG) signals recorded from 60 subjects have been classified as ‘normal’ or ‘at risk’. In this process, we have used an Auto-Regressive eXogenous input (ARX) linear parametric model for extracting features that represent the circulatory system and a support vector machine (SVM) for classifying the signals based on the four data segment selection policies; best fit, three best fit, ten best fit and average best fit. The classification method employed in this work appears to be novel. According to the sensitivity and the specificity obtained (84.615% and 92.31%, respectively), the average best fit policy was chosen as the best policy for the classification of PPG signals.

Important Links:

Go Back