Dear Hamid, > Am 22.03.2017 um 17:09 schrieb Hamid Ghofrani <ghof...@gmail.com>: > > Dear All, > We collect additional entries from dbpedia live from time to time. I noticed > the number of entries has shrunk.
What do you mean by „entries“, rdf:type-relationships? > For example for Actor , we used to get about 40K actors from live updates > (http://dbpedia.org/ontology/Actor) > > now it is just down to 6500. This might be because Wikipedia editors tend to use more generic templates, e.g. Person instead of Actor. DBpedia uses these templates to map the resource to a class from the DBpedia ontology. This is truly an issue, a lot of interesting stuff is removed from infoboxes in a concerted way, e.g. notable works, influenced persons, and more. It would be interesting to check, whether (and which) relevant information from past versions could be perpetuated to be continuously part of current releases. > Another issue regarding entries > We used to get > <http://dbpedia.org/resource/Tom_Cruise> > <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> > <http://dbpedia.org/ontology/Actor> > > from live Updates. It does not exist anymore, why? I cannot say when this was the case. https://en.wikipedia.org/wiki/Tom_Cruise uses the Person template at least since January 2013 [1]. So it hardly should have been typed an actor since then. Nevertheless, this type could have been part of the releases, introduced by data cleansing and type inference post processing steps. > Also why previous liveUpdate results were not added to the 2016 release? DBpedia Live and the Releases are not published together and results are not mixed. DBpedia Live extracts the latest state from Wikipedia as soon as the article changes, so there may be little date time differences between the extraction times of individual articles. The releases on the opposite side are all based on a single dump from Wikipedia, i.e. all having the same extraction time. Furthermore, the release cycle allows to do more complex extractions, e.g. internationalization, page links, and post processing steps such as type inference, page rank, etc. These are not feasible with the constant flux date nature of DBpedia Live. > Thank you very much Best regards Magnus [1] https://en.wikipedia.org/w/index.php?title=Tom_Cruise&action=edit&oldid=530693566 -- Magnus Knuth Hasso-Plattner-Institut für Softwaresystemtechnik GmbH Prof.-Dr.-Helmert-Str. 2-3 14482 Potsdam Amtsgericht Potsdam, HRB 12184 Geschäftsführung: Prof. Dr. Christoph Meinel tel: +49 331 5509 547 email: magnus.kn...@hpi.de web: http://www.hpi.de/ webID: http://magnus.13mm.de/ ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ DBpedia-discussion mailing list DBpedia-discussion@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
