Date of Award
Master of Public Health (MPH)
This study investigates three alternative machine learning methods to explore influential predictors of type 2 diabetes. It compares ridge, lasso, and elastic net regression to linear regression, and focuses on 12 outcome variables that include age, sex, race, income, education level, body mass index, waist circumference, arm circumference, hip circumference, family history, smoking status, sleep duration, high blood pressure, and high-density lipoprotein. Ridge, lasso and elastic net regression do not outperform linear regression but do assist in choosing a simpler model which could be important for improving future modeling.
Chernishkin, Amanda E., "Identifying Influential Variables in the Prediction of Type 2 Diabetes Using Machine Learning Methods." Thesis, Georgia State University, 2020.
File Upload Confirmation