A Comparative Assessment of Synthetic Data Augmentation Methodologies (SMOTE, Variational Autoencoders, and GANs) in Resolving Class Imbalance for Predictive Diabetes Analytics. TSP 2026, 2 (2).