Graduation Semester and Year
Spring 2026
Language
English
Document Type
Thesis
Degree Name
Master of Science in Computer Science
Department
Computer Science and Engineering
First Advisor
Dr. Jacob Luber
Second Advisor
Dr. Jenniffer Woo
Abstract
Preterm birth (PTB), defined as delivery before 37 completed weeks of gestation, is a complex condition with overlapping, interacting biological pathways. Many studies have looked at predictive modeling of PTB using classical machine learning models. Most of these studies rely on only a single or very few features to train their models. Some are strictly clinical, while others study genomic or other omics data. A common gap, however, is that most studies do not account for the disproportionate impact on certain demographics; specifically, non-Hispanic Black women have been linked with a significantly higher risk of PTB. Our study tries to bring together multiple biological pathways and risk factors for PTB in a cohort of non-Hispanic Black women (N = 184). We analyze model interpretability using the SHapley Additive exPlanations (SHAP) framework and present our results. We are able to verify existing literature and also identify features not previously associated with PTB.
Keywords
Interpretability, Feature Attribution, Mutli-omics Integration, Biomarkers, Pathways, Health Disparities, Precision medicine
Disciplines
Computational Biology | Computer Sciences | Maternal, Child Health and Neonatal Nursing
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
Recommended Citation
Yumkhaibam, Budhachandra Singh, "Multi-Modal Modeling of Preterm Birth Risk in an African-Ancestry Cohort" (2026). Computer Science and Engineering Theses - Archive. 545.
https://mavmatrix.uta.edu/cse_theses/545
Included in
Computational Biology Commons, Computer Sciences Commons, Maternal, Child Health and Neonatal Nursing Commons
Comments
This work could not have been possible without the support of everyone. Thank You!