Quality Scoring
The Scoring Engine aggregates individual detections into a unified quality score and decision.
Quality Dimensions
ADQA categorizes every detection into one of several standard quality dimensions:
| Dimension | Description |
|---|---|
| Validity | Conformity to logical rules and semantic ranges. |
| Completeness | Presence of required values (null-free). |
| Uniqueness | Freedom from redundancy and duplicates. |
| Consistency | Statistical stability and lack of drift. |
| Privacy | Absence of sensitive or non-compliant PII. |
Scoring Logic
- Detection Scoring: Each issue is assigned a severity score (0.0 to 1.0).
- Dimension Aggregation: Scores are aggregated within each dimension using a weighted average or maximum severity.
- Global Score: A final quality score is calculated for the entire dataset.
Quality Decision
Based on the global score and configured thresholds, ADQA makes one of three decisions:
- ✅ PASS: No critical issues found. Global score is above the "Pass" threshold.
- ⚠️ WARN: Potential issues detected. Score is below "Pass" but above "Fail".
- ❌ FAIL: Critical quality risks identified. Score is below the "Fail" threshold.