Monitoring patient's condition based on breath detection

Abstract

A constant increase in elderly population in developed countries closely followed by a continuous reduction of costs, inter alia in healthcare, calls for inventing efficient methods in eldercare. For that purpose, this study is devoted to an introductory elaboration of a computer-based monitoring system that extracts patient information based on respiration analysis. By combining the field of respiratory medicine with machine learning, an empirical study has been conducted in accordance to prior methodical review of the state-of-the-art research. Consequently, Sleep Apnea-Hypopnea Syndrome is explored in the context of automated event classification by applying Artificial Neural Network and Support Vector Machine classifiers. Experiments have not resulted in revolutionary findings, however this thesis contributes with several valid suggestions and may be used as an introduction to the field and a foundation for further research.