QAVCM
Carbon Markets
Methodology / Mapping Methodology

Mapping Methodology

How cross-registry entity matching works — features, confidence model, merge logic, and manual overrides.

Live config
Loading live config…
Examples
VCS-191↔carbonmark:VCS-191
Auto-merge
94%
identifier
100+20
name
92+18
country
100+20
scheme
85+17
vintage
95+19
Exact identifier overlap + high name similarity → above auto-merge threshold.
GS-2061↔cad:GS-2061-variant
Review queue
75%
identifier
60+12
name
78+16
country
100+20
scheme
90+18
vintage
70+14
Partial identifier match, name drift, vintage mismatch → falls in review band.
VCS-9999↔regen:unknown-project
Discard
28%
identifier
0+0
name
42+8
country
50+10
scheme
30+6
vintage
20+4
No identifier overlap, low name similarity, scheme mismatch → below discard threshold.
Logic
PhaseScore rangeBehaviour
Soft candidateAny score > 0Pair is computed and stored. No action taken — used for explainability only.
Review queue≥ review thresholdPair is surfaced in Mapping Review for a human decision. Can be merged or dismissed.
Auto-merge≥ autoMerge thresholdEntities are merged automatically. Mapping decision is logged with score + feature breakdown.
Manual overrideAnyOperators can force-merge or force-dismiss any pair regardless of score.