Kavitha Srinivas wrote:
> In case anyone is interested, we tried to statistically map freebase
> types to dbpedia types, the key ideas are described here:
> http://domino.watson.ibm.com/library/cyberdig.nsf/papers/
> 4D84639C32795569852574FD005EA539/$File/rc24684.pdf
> A variant of this paper was just accepted to the International
> Semantic Web Conference (ISWC).
> Thanks!
>
These guys have been a bit more systematic that I have, but I've
used pretty much the same method in my type reconciliation efforts.
Lots of interesting things fall out of this kind of effort. For
instance, I was trying to separate out "Organizations" (In the dbpedia
sense, which is wider than the Freebase sense...) from "(Creative)
Works" because my next processing stage works for one and not for the other.
Tracking inconsistencies between FB and the Dbpedia Ontology, I
discovered that a "Newspaper", which is considered a "Work" by the
Dbpedia Ontology, is also an "Organization" (although the Dbpedia
Ontology doesn't know this.) This can be established by "duck typing",
ex.,
"ADAM LIPTAK
<http://topics.nytimes.com/top/reference/timestopics/people/l/adam_liptak/index.html?inline=nyt-per>
works for the New York Times" is a good assertion, but you can't say
that "Issac Asimov works for The Foundation Trilogy."
------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Dbpedia-discussion mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
