Photoplethysmographic Waveforms Analysis and Classification

Abstract

The work is devoted to the automatic classification of plethysmographic signals. The application of machine learning methods to classify plethysmographic signals has been studied. Combined use of k-means and agglomerative clustering methods for classifying pulse waves according to morphological types is proposed. The methods of signal preprocessing are considered. The optimal combination of features is estimated, and methods for their selection are considered. An automatic pulse wave classifier has been obtained that does not require annotated training samples. The results of a computer experiment on the automatic classification of signals are presented.