Utarbeidet med KI.
Why is identity matching difficult?
Determining whether two records refer to the same person sounds simple, but rarely is. Names are spelled differently, addresses are abbreviated, people move, and companies merge.
Spouses who share an email address. Business addresses used as home addresses. Names with and without middle names. And in Norway in particular: the same company name in multiple municipalities, distinguished only by the organization number.
If two different customers are merged, one will be able to see the other’s history. At best, this is embarrassing; at worst, it is a breach of privacy. Leaving two profiles unmerged is inconvenient; merging the wrong profiles is serious. Set the threshold accordingly.
For business customers, the organization number is the only truly unique identifier. If it is available in the source data, matching should be based on it rather than name similarity. If it has not been recorded, adding it is often the most worthwhile data-cleaning task you can undertake.
Still et spørsmål eller del hva som hjalp deg.
Del et spørsmål eller en refleksjon.
Bli den første som bidrar.