A Supervised Learning Framework for Predicting GSC Antibody Seropositivity in Guillain–Barré Syndrome Using Multivariate Clinical and Demographic Indicators
Authors:-Ms. Sangeetha Raj S, Ayushi Negi, Ekta Kumari, Christina S
Abstract-:This paper proposes a robust supervised learning framework for predicting ganglioside complex (GSC) antibody seropositivity in patients with Guillain–Barré Syndrome (GBS) using multivariate clinical and demographic features. Drawing from a comprehensive dataset encompassing 129 GBS patients, we employed advanced machine learning methods support vector machines, random forests, decision trees, and k-nearest neighbours to predict seropositivity for six key anti-ganglioside antibodies (GM1, GM2, GD1a, GD1b, GT1b, GQ1b). Rigorous feature selection, cross-validation, and class imbalance handling were implemented to ensure robustness. Results show that routine clinical data can deliver accurate antibody seropositivity predictions, supporting GBS management where serological assays are delayed or unavailable.
