ic_006 · Industry: Manufacturing — QC failure prediction
Status: ✓ Passing · Last run: 2026-06-07T07:00:52 · Pipeline: v2
· History: 1/1 runs passed (100%)
What was tested
Manufacturing — QC failure prediction Defect prediction: units + inspections + sensor_readings + labels. Signal: anomalous vibration/temperature + inspection failures.
What we planted in the data
- 300 entities
- Signal strength: 0.70 (sigmoid slope multiplier)
What the system did
- Training rows built: 1,200
- Features used in model: 23
- Model selected:
RandomForestClassifier - CV AUC: 0.774
Top features by importance:
vibration_avg_365d(0.231)vibration_avg_90d(0.146)temperature_max_180d(0.083)vibration_sum_90d(0.081)days_since_production(0.074)
What we checked — all assertions passed
| Status | Assertion | Detail |
|---|---|---|
| ✓ | no_pipeline_errors |
clean run |
| ✓ | feature_survival_rate_above |
23/23 features survived (100% ≥ 40%) |
| ✓ | source_tables_contributing_above |
7 distinct source-table prefixes in surviving features (['days', 'inspection', 'model', 'production', 'shift', 'temperature', 'vibration']) |
| ✓ | auc_in_band |
AUC = 0.774 inside band [0.55, 0.95] |
How this could have gone wrong
(no assertion descriptions provided)
Why this case matters
Industry-canonical ML problem; multi-table shape mirrors real customer datasets.
Reproducing
# from auto_insight_api/
python -m validation.v2 run ic_006 --pipeline v2 -v
- Case config:
validation/v2/cases/ic_006_manufacturing_defect.yaml - Data shape:
industry_manufacturing_defect - Analytics type:
ml_binary