Document Type

Thesis

Degree Name

Master of Science (MSc)

Department

Mathematics

Faculty/School

Faculty of Science

First Advisor

Xu (Sunny) Wang

Advisor Role

Associate Professor

Second Advisor

Yang Liu

Advisor Role

Professor

Abstract

Sleep apnea is a prevalent and potentially serious sleep disorder, affecting 1% to 6% of children and adolescents [1]. Early and accurate diagnosis is critical, as untreated sleep apnea can lead to severe health complications, including cardiovascular problems and developmental issues. Traditional diagnostic methods often rely on manual analysis of polysomnography (PSG) data, which can be time-consuming and require specialized expertise. This thesis explores the use of Fourier transforms, differencing, and other phys- iological signal processing techniques to develop a transformer-based diagnostic model from scratch, utilizing multi-dimensional time series data from PSG. Our research distin- guishes itself by not only comparing conventional diagnostic methods with cutting-edge transformer architectures but also by aiming to automate the diagnostic process. The focus is on creating a model that can accurately diagnose obstructive sleep apnea (OSA) with- out extensive domain knowledge, thus making the diagnostic process more accessible and efficient. The findings underscore the significance of robust data processing and model selection in enhancing diagnostic accuracy, potentially revolutionizing the field of sleep medicine.

Convocation Year

2024

Convocation Season

Fall

Download

Share

COinS