Datasets:
| { | |
| "READY": true, | |
| "name": "CDC_Diabetes_Health_Indicators", | |
| "task": "binary", | |
| "num_classes": 2, | |
| "cat_idx": [ | |
| 7, | |
| 8, | |
| 9, | |
| 10, | |
| 11, | |
| 12, | |
| 13, | |
| 14, | |
| 15, | |
| 16, | |
| 17, | |
| 18, | |
| 19, | |
| 20 | |
| ], | |
| "num_samples": 226342, | |
| "num_features": 21, | |
| "imbalance_ratio": 5.7502, | |
| "description": "Talent Benchmark dataset: CDC_Diabetes_Health_Indicators (task=binary). Rows from official train/val/test files are concatenated; TabBioMed then applies a 70/15/15 train/val/test split (stratified for classification, random_state=42), not the original Talent split boundaries. Exact duplicate rows were collapsed, and feature patterns with conflicting labels were dropped.", | |
| "source": "Talent Benchmark Dataset - CDC_Diabetes_Health_Indicators", | |
| "label": "clinical & healthcare" | |
| } |