Performance Analysis of Least Square Linear Regression with Various Classifiers for Cardiovascular Respiratory Detection from Capnography

Abstract

In this study, the PPG signal was taken from a capnography data source and used along with statistical characteristics and machine learning techniques to diagnose respiratory disorders. After the statistical properties of the data have been extracted using a method called least square linear regression, the signal is then processed by a number of different classification tasks, and the outcomes of the classification algorithm are examined. Results show that the linear regression and nonlinear classifiers gives the best accuracy of 88.11% and 85.73% for both normal and respiratory disease cases.