Date of Award
8-2024
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Industrial Engineering
Committee Chair/Advisor
Hamed Rahimian
Committee Member
Amin Khademi
Committee Member
Qi Luo
Abstract
In modern healthcare, the convergence of data-driven methodologies and decision making frameworks has become imperative for enhancing patient-care, operational efficiency, and resource allocation. This thesis explores the synergistic integration of two powerful paradigms: machine learning (ML) and Markov decision processes (MDPs), to address key challenges in healthcare management. Chapter 2 showcases the integration of ML techniques for predicting laboratory results using Electronic Medical Records (EMR) data. By harnessing the information embedded within EMRs, ML models offer predictive insights into patient outcomes, facilitating early detection of health risks and enabling personalized interventions. Various ML algorithms suitable for handling EMR data are discussed, alongside challenges related to data preprocessing and model interpretability. Furthermore, recent advancements in deep learning architectures are explored, showcasing their efficacy in longitudinal predictions and clinical decision support. Chapter 3 shifts focus to MDP-based policy analysis, emphasizing its utility in optimizing healthcare decision-making processes. Fundamental concepts of MDPs, algorithms for solving them, and extensions to accommodate real-world complexities are examined. Through implementations and numerical experiments, the chapter illustrates the practical implications of integrative analytics in healthcare settings, highlighting their potential to revolutionize patient-care, resource allocation, and policy development.
Recommended Citation
Harrison, Charles, "Personalized Treatment Recommendations for Diabetes Patients" (2024). All Theses. 4384.
https://open.clemson.edu/all_theses/4384