Performance of the four trained SNNN models contrasting classification metrics k-fold cross-validation (k=10) of the SNNN algorithm and the performance in the testing (NHANES 1999–2004) and training datasets (NHANES-III)
| Model | Cohort | Accuracy | AUC SIRD (95% CI) | AUC SIDD (95% CI) | AUC MARD (95% CI) | AUC MOD (95% CI) |
| HOMA2-IR, HbA1c, age of diabetes onset, HOMA2-β, BMI | Training | 0.986±0.004 | 1.00 (0.99 to 1.00) | 1.00 (1.00 to 1.00) | 1.00 (1.00 to 1.00) | 1.00 (0.99 to 1.00) |
| Validation | 0.981±0.008 | 1.00 (1.00 to 1.00) | 0.99 (0.99 to 1.00) | 1.00 (0.99 to 1.00) | 1.00 (1.00 to 1.00) | |
| HOMA2-IR’, HbA1c, age of diabetes onset, HOMA2-β’, BMI | Training | 0.894±0.008 | 0.97 (0.95 to 0.97) | 0.99 (0.99 to 1.00) | 0.98 (0.97 to 0.99) | 0.99 (0.99 to 1.00) |
| Validation | 0.903±0.008 | 0.86 (0.82 to 0.90) | 0.97 (0.95 to 0.99) | 0.88 (0.85 to 0.91) | 0.93 (0.91 to 0.93) | |
| HOMA2-IR, glucose, age of diabetes onset, HOMA2-β, BMI | Training | 0.857±0.007 | 0.96 (0.95 to 0.97) | 0.98 (0.98 to 0.99) | 0.97 (0.96 to 0.98) | 0.99 (0.98 to 0.99) |
| Validation | 0.859±0.007 | 0.84 (0.80 to 0.88) | 0.85 (0.81 to 0.89) | 0.84 (0.81 to 0.87) | 0.92 (0.89 to 0.95) | |
| HbA1c, age of diabetes onset, BMI, METS-VF, METS-IR | Training | 0.810±0.006 | 0.89 (0.87 to 0.91) | 0.99 (0.99 to 1.00) | 0.95 (0.94 to 0.96) | 0.98 (0.98 to 0.99) |
| Validation | 0.820±0.006 | 0.82 (0.78 to 0.86) | 0.95 (0.93 to 0.98) | 0.88 (0.85 to 0.90) | 0.94 (0.92 to 0.97) |
AUC, area under the receiver operating characteristic curve; BMI, body mass index; HbA1c, glycated hemoglobin; HOMA, homeostasis model assessment; IR, insulin resistance; MARD, mild age-related diabetes; METS-IR, metabolic score for insulin resistance; METS-VF, metabolic score for visceral fat; MOD, mild obesity-related diabetes; NHANES, National Health and Nutrition Examination Survey; SIDD, severe insulin-deficient diabetes; SIRD, severe insulin-resistant diabetes; SNNN, self-normalizing neural network.