Publication Date

2026

Document Type

Dissertation/Thesis

First Advisor

Ryu, Duchwan

Degree Name

M.S. (Master of Science)

Legacy Department

Department of Statistics and Actuarial Science

Abstract

This thesis investigates symbolic logistic regression for interval-valued predictors through simulation studies and a real health data application. Classical logistic regression assumes exact predictor values, whereas in many practical settings variables are available only in interval form due to coarsening or reporting uncertainty. Two simulation studies examine the impact of interval uncertainty under asymmetric intervals and measurement error. Symbolic models based on midpoint and midpoint-plus-width representations are compared with the classical approach. Results show that midpoint modeling captures the general relationship but introduces bias under asymmetry, while incorporating width reduces this distortion. Under measurement error, classical logistic regression exhibits attenuation bias, whereas symbolic models remain comparatively stable. The methodology is applied to NHANES data using grouped age intervals, where symbolic midpoint regression closely reproduces classical fitted curves with only modest attenuation. Overall, symbolic logistic regression offers a stable and interpretable approach when predictors are available in interval form.

Extent

46 pages

Language

en

Publisher

Northern Illinois University

Rights Statement

In Copyright

Rights Statement 2

NIU theses are protected by copyright. They may be viewed from Huskie Commons for any purpose, but reproduction or distribution in any format is prohibited without the written permission of the authors.

Media Type

Text

Download

Share

COinS