Identifying and assessing the severity of Acute Respiratory Distress Syndrome with machine learning methods
Name
1227274737-MIT.pdf
Size
1.07 MB
Format
Adobe PDF
Checksum (MD5)
7fd161aa5237607c9977d4ddc2bd0c48
Author(s)
Boyer, Yun(Yun X.)
Advisor(s)
Alistair Johnson.
Date Issued
2020
Publisher
Massachusetts Institute of Technology
Abstract
Acute Respiratory Distress Syndrome (ARDS) is a respiratory failure wherein alveoli become filled with excess fluid; it can be life-threatening. Worldwide recognition/ identification of ARDS is as low as 51.3%. Therefore, there is a need for better methods for its diagnosis, and machine learning methods may offer a solution. To increase consistency amongst ARDS diagnoses, an accurate quantification system can be built to leverage all available information sources regarding the disease. For example, sources such as electronic hospital records (EHR) and X-ray images can be used to train models for this qualification system. Such a system would increase consistency amongst ARDS diagnoses and would help with the understanding of the disease by allowing better comparisons among cases of ARDS. This project shows that numerical features provides predictive information and can predict the mortality of ARDS patients with AUROC of .75 on the never-seen testing set. However, it is inconclusive whether or not X-rays can provide additional information as the dataset was too small to train all the parameters of the computer vision model.
Description
Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, September, 2020
Cataloged from student-submitted PDF of thesis.
Includes bibliographical references (pages 35-36).
Subjects
Electrical Engineering and Computer Science.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
MIT theses may be protected by copyright. Please reuse MIT thesis content according to the MIT Libraries Permissions Policy, which is available through the URL provided.
Persistent DSpace Link
