> > It's a precaution: Wikipedia categories are a navigational aid, and can't > > be used reliably to find things of a certain kind. > 1. apply a FILTER based on your known > knowledge (e.g., what you've pointed out above) > 2. enhance this list with your domain knowledge > and then hopefully publish back to the LOD cloud
Hi Kingsley! The problem is: 1. How to KNOW which categories represent types. There are 5-10 approaches for this, based on NLP and ML. E.g. Yago2 assumes that if the head-word of the cat name is a plural noun, that's a class. MENTA improves this by looking for a head-word that's a countable noun. A lot of them tie up into Wordnet, some into OpenCyc/UMBEL. It's a hard research problem. 2. How to KNOW which category-entity instance is an exception. E.g. a category "Books of Author Xyz" is typically applied to the page "Author Xyz" and a naïve interpretation will conclude that the author is a book. In wikipedia, Categories are a navigational aid, i.e. "mere" links. The trick is how to find those links that are type links. > >> >http://dbpedia.org/c/9CH4UXVL -- shortened version of the above > > Kingsley, how do you get this? My question was about the shortening. Hacking the URL I find that the dbpedia URL shortener is at http://dbpedia.org/c/ But is there a quick way to invoke it from SPARQL results? Should I make a browser bookmark? ------------------------------------------------------------------------------ Dive into the World of Parallel Programming! The Go Parallel Website, sponsored by Intel and developed in partnership with Slashdot Media, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials and more. Take a look and join the conversation now. http://goparallel.sourceforge.net _______________________________________________ Dbpedia-discussion mailing list Dbpedia-discussion@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
