Diabetes Subtype Clustering

The clustering analysis selected k = 2 for clinical interpretability despite only modest separation. Cluster 0 aligned with Type 1 diabetes phenomenology because it had lower insulin and younger age. Cluster 1 aligned with Type 2 diabetes phenomenology because it had higher insulin and older age. The subtype clustering is exploratory and does not establish a validated Type 1 or Type 2 classifier. The Type 1 and Type 2 assignment lacked ground-truth labels and was inferential. Stage 2 clustered confirmed diabetic cases using standardized glucose, insulin, and age.