A Network Approach To Predict GDM Risk In Pregnant Women

Abstract

High blood sugar during pregnancy known as gestational
diabetes mellitus (GDM) can cause difficulties for
both the mother and the unborn child. Particularly in
places where prenatal care is scarce, early detection and
management are essential. This study suggests a combined
machine learning prediction model to determine which
expectant mothers are susceptible to gestational diabetes
mellitus. We examined eight distinct models, incorporating
deep learning methodologies.(Artificial Neural Networks)
and conventional machine learning algorithms (Support
Vector Machine, Naive Bayes, Random Forest, and Logistic
Regression), using a dataset of 3526 pregnant women
from Kaggle’s Gestational Diabetes Mellitus dataset. With
accuracy rates ranging from 87% to 97%, these models
demonstrate the immense potential of machine learning to
enhance GDM screening and management, especially in
resource-constrained environments.