Explain the theories behind the AI/ML solution you develop in a case study or take-home challenge. For this SQL implementation, produce an evaluation summary from model runs, datasets, and prediction results. Include classification metrics, comparison with the prior run for the same model, and relative performance among runs evaluated on test data.
run_id, model_name, dataset_name, evaluated_cases, precision, recall, f1_score, previous_f1_score, f1_change, and performance_rank.dataset_name, then run_started_at ascending.| Column | Type | Description |
|---|---|---|
| dataset_idPK | INT | Dataset identifier |
| dataset_name | VARCHAR(100) | Dataset name |
| split | VARCHAR(20) | Dataset split |
| Column | Type | Description |
|---|---|---|
| run_idPK | INT | Model run identifier |
| model_name | VARCHAR(100) | Model name |
| dataset_id | INT | Evaluated dataset identifier |
| run_started_at | DATE | Date the run started |
| hypothesis | VARCHAR(200) | Short modeling hypothesis |
| Column | Type | Description |
|---|---|---|
| prediction_idPK | INT | Prediction identifier |
| run_id | INT | Model run identifier |
| entity_id | VARCHAR(50) | Scored entity identifier |
| actual_label | BOOLEAN | Observed binary label |
| predicted_label | BOOLEAN | Predicted binary label |
| confidence | DECIMAL(5,4) | Prediction confidence |