University of North Florida. Department of Mathematics and Statistics
A comparative study of classification methods for healthcare analytics
Abstract
dc:description.abstractThis thesis presents a comparative study of logistic regression, Linear Discriminant Analy- sis (LDA), and Quadratic Discriminant Analysis (QDA) for binary classification in healthcare analytics, integrating theoretical derivation, simulation, and real-data application. A facto- rial simulation study crosses the covariance structure (equal vs. unequal), predictor correla- tion (ρ ∈ {0, 0.5, 0.9}), dimensionality (p ∈ {2, 5, 10}) and sample size (n ∈ {50, 100, 200}) across 54 scenarios with 1,000 Monte Carlo replicates each. Three main findings emerge. Logistic regression and LDA are nearly interchangeable when the assumption of equal-covariance holds. QDA achieves substantially better discrimi- nation when class-specific covariance matrices differ, with AUC gains of 0.05–0.07. However, QDA’s variance penalty makes it unreliable when sample sizes are small relative to dimen- sionality. These findings are validated using NHANES 2017–2018 diabetes data (n = 3,703), where the survey-weighted logistic regression achieves a concordance statistic of 0.806 and LDA outperforms QDA in the cross-validated error rate (0.287 vs. 0.300). The results provide practical guidance for classifier selection in healthcare settings based on covariance structure, sample size, and inferential goals.
Degree
thesis:*- Name thesis:degree_name
- Master of Science in Mathematical Sciences (MS)
- Discipline thesis:degree_discipline
- Mathematics & Statistics
- Grantor
- University of North Florida. Department of Mathematics and Statistics
- Date dc:date.issued
- Spring
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Zhao, Xueting
- Zhao, Xueting, 1997-
- Advisors dc:contributor.advisor
-
- Jia, Yisu
- Xu, Jasper
- Wang, Peiyao
- Contributors dc:contributor
-
- Dr. Yisu Jia
- Dr. Jasper Xu
- Dr. Peiyao Wang
Subjects
dc:subject × 14- Thesis
- University of North Florida
- UNF
- Dissertations, Academic -- UNF -- Master of Science in Mathematical Science
- Dissertations, Academic -- UNF -- Mathematics
- classification, logistic regression, linear discriminant analysis, quadratic discriminant analysis, healthcare analytics, NHANES
- Applied Statistics
- Biostatistics
- Statistical Methodology
- Statistical Models
- Statistical Theory
- Statistics and Probability
- Vital and Health Statistics
- Medical informatics -- Classification -- Evaluation; Medical care -- Classification -- Evaluation; Diabetes -- Data processing -- Evaluation; Health surveys -- Data processing -- Evaluation; Classification -- Data processing -- Technological innovations; Discriminant analysis; Logistic regression analysis
Rights
dc:rightsIdentifiers
dc:identifier.*- Repository record dc:identifier
- https://digitalcommons.unf.edu/etd/1419