Closed SvenLieber closed 3 months ago
That overlap comes directly from the manually curated correlation list. We should further investigate the evolution of those correlation lists to exclude that by accident a copy-paste Excel operation falsified the data.
Apparently some persons from Unesco Index Translationum are integrated but have nothing in common. For example
4431
u1803
affef438-2ca2-4f1c-b5e6-535343772db9
u1803
The unesco ID is not part of the Unesco data source, but was created by us by initially grouping similar names. However, over time we also started to use a BELTRANS corpus-wide correlation list of persons, also a possible source of errors