Predicting Gestational Diabetes and PCOS Susceptibility: A Machine Learning-Based Cross-Framework Analysis

Polycystic ovarian syndrome (PCOS) and gestational diabetes mellitus (GDM) are significant health concerns that affect pregnant women, often leading to adverse pregnancy outcomes. Early and accurate prediction of these conditions is crucial for timely intervention and management. This study conducts a comprehensive cross-framework analysis using multiple machine learning (ML) models to predict GDM and assess the likelihood of PCOS in pregnant women. We used four benchmark data sets containing key health parameters such as glucose levels, body mass index (BMI), insulin levels, family history, and hormonal indicators. ML models, including Logistic Regression (LR), K-Nearest Neighbor (KNN), Decision Tree Classifier (DTC), Gradient Boosting (GB) are used for their ability to detect complex health patterns and improve early & accurate disease prediction and neural networks like Multi-Layer Perceptron (MLP), FeedForward Neural Network (FNN), and Convolutional Neural Network (CNN) are used for their ability to capture complex, non-linear patterns in high- dimensional medical data. We achieved 93% accuracy using Gradient Boosting (GB), FNN, and CNN on Diabetes_dataset_1, and KNN, MLP, FNN, and CNN on Diabetesdataset2. Neural networks demonstrated superior performance across both datasets, highlighting their effectiveness in predicting outcomes.