Estimating the Risk of a Down's Syndrome Term Pregnancy Using Age and Serum Markers: Comparison of Logistic Regression, Quadratic Discriminant Analysis, and Support Vector Machines

Abstract

This thesis aims to enhance the precision and safety of Down's syndrome (DS) risk assessment by exploring innovative, non-invasive approaches in prenatal healthcare. Traditional invasive procedures, like amniocentesis, pose risks, motivating our investigation into alternative methods. We concentrate on Logistic Regression, Quadratic Discriminant Analysis (QDA), and Support Vector Machines (SVM), each offering unique strengths in risk estimation.