
Researchers have developed an AI-powered tool capable of classifying diabetes into four specific diagnostic categories: prediabetes, type 1, type 2, and pancreatogenic diabetes.
The model was trained and tested on a multiclass dataset of 21,539 samples, aiming to improve patient outcomes by enabling more personalized treatment plans based on accurate diagnostic labels.
While the tool offers potential for rapid screening in primary care, experts note that clinical validation and patient data privacy remain critical challenges for its integration into healthcare systems.