Modeling Human-Informed Variables in Medical Data
Name
abudaoud-georgead-meng-eecs-2025-thesis.pdf
Description
Thesis PDF
Size
964.48 KB
Format
Adobe PDF
Checksum (MD5)
844493475d927f2827bad0b7cd447013
Author(s)
Abu Daoud, George
Advisor(s)
Ghassemi, Marzyeh
Date Issued
May 2025
Publisher
Massachusetts Institute of Technology
Abstract
In the Age of Information and Artificial Intelligence, data plays a major role in analyzing and understanding underlying trends and patterns as well as informing processes and operations. Medical data often captures information beyond mere patient conditions and state, but also human behavioral aspects of the medical process, affecting the data itself and the decisions informed by it. Modeling these variables could help us understand how they influence decisions in the field and potentially augment our models for better and more nuanced predictions. In the first study, we look into how external non-medical factors might affect decision-making by investigating the effect of 30-day mortality metrics on discharge rates following surgeries in Cardio-Vascular Intensive Care Units (CVICU) using data from the MIMIC-IV dataset. In the second study, we examine data extraction from human-notes for enhancing organ procurement decision processes.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
In Copyright - Educational Use Permitted
Copyright retained by author(s)
Persistent DSpace Link
