We test whether the model remains
conceptually sound, data-appropriate, statistically reliable, and fit
for its intended use.
Model dataÂ
collection
01
02
Data cleansing
Default-data mapping
03
04
Data consolidation
05
Data analysis
Validation workflow
Availability of appropriate consolidated validation data
Two Complementary Assessments
Qualitative Validation
Is the model appropriately designed, governed and supported?
-
Rating-system designÂ
-
Model governanceÂ
-
Data qualityÂ
-
BenchmarkingÂ
Qualitative Validation
Does observed model performance support continued use?Â
-
Model stabilityÂ
-
Predictive-power validationÂ
-
Discriminatory-power validationÂ
-
Back-testingÂ
-
Model granularityÂ
Validation documentation
Clear evidence, findings and limitations
Targeted remediation where performance gaps emerge
Model modification
Governance feedbackÂ
Input to approval, monitoring and periodic review
Outcomes & Continuous Improvement