Re: [Dbpedia-discussion] Advancing the DBpedia ontology

Hi Alexandre,

In the field of Linked Data we cannot only talk of ontology matching or
instance matching as separate fields, in my opinion its something in the
middle, as you mostly have to check predicates, properties and class but
also match the instance values if you need to interlinking resources.

I'm talking of using instance information and matching distances to create
owl:sameAs interlinking. For example,  geo:lat, geo:long can be used in
order to match (e.g. using haversine distance) with other geolocated
instances, or string matching measures like Jaro-Winkler could be applied
for rdfs:label or any other string label.


Some useful references:
Volz, Julius, et al. Discovering and maintaining links on the web of data.
Springer Berlin Heidelberg, 2009.

Kalpa Gunaratna, Sarasi Lalithsena andAmit Sheth. Alignment and dataset
identification of linked data in Semantic Web. DOI: 10.1002/widm.1121

Szekely, Pedro, et al. "Connecting the Smithsonian American Art Museum to
the Linked Data Cloud." The Semantic Web: Semantics and Big Data. Springer
Berlin Heidelberg, 2013. 593-607.

Hope this helps

Feel free to contact me

Antonino


> Hi Dimitris and Antonino,
>
> A couple of questions:
>
> When you mentioned interlinking methods, are you talking about instances
> matching or schema matching? I am very interested in methods for instances
> deduplication in a triplestore. Any reference?
>
> How do you keep a version control of the ontology changes with WebProtege?
> Does it work fine with many contributors editing the ontology
> simultaneously? How does DBPedia ontology evolved over time? These are all
> relevant questions for other community-driven ontologies like VIVO (
> http://vivoweb.org).
>
> Best,
>
> Alexandre Rademaker
> http://arademaker.github.com/
>

Received on Saturday, 24 January 2015 12:40:57 UTC