Links to DBpedia added

3 december, 2008 Martin

Just added 4.669 links to DBpedia. Just persons for now. I used the german persondata dataset to make the mapping. The assumption I am making is that if the string [name] + [year of birth] maps to exactly one URI in each database, then they are assumed to be the same person. This does not work 100%, on the other hand nothing does. What’s nice though is that I get to actually use the different names in the authority record.

However, since DBpedia has some books as well the mapping could be made more exact by matching our books to theirs. In the future, given that more libraries embrace Linked Data, disambiguating authors with the same name should be fairly easy when you have access to some of their books as well. I’m guessing that is what is happening behind the scenes at VIAF.

There’s more coming, for example persons with non-ascii characters in their names did not match due to differences in encoding. In the end we will have to do some manual work as well. Quality control and names that needs to be disambiguated, etc, etc.

Example:
Jimi Hendrix – http://libris.kb.se/resource/auth/243497

Filed under: Metadata

Etiketter:,


Kategorier

Senaste kommentarer

Arkiv

Verktyg

Informationskanaler