1

I'm trying to use dbpedia spotlight to spot special terms (which is not included in dbpedia) by using a local mediawiki dump as an input instead of the default index and spotter.dict. Any ideas will be so appreciated

Toby Allen
  • 10,997
  • 11
  • 73
  • 124
mOgaio
  • 43
  • 4
  • I tried to simulate what dbpedia do with wikipedia locally but in stage i stopped that because of the lake of the hardware resources. I will be so please if you have any ideas about how can i create an index and spot.dict out of the mediawiki to use in dbpedia spotlight. – mOgaio May 14 '13 at 09:51
  • There are lots of instruction on the [DBpedia spotlight wiki](https://github.com/dbpedia-spotlight/dbpedia-spotlight/wiki). How far did you get in installing and setting everything up? – Joshua Taylor Jun 24 '13 at 11:12

1 Answers1

1

DBpedia Spotlight requires 5(five) files to build the index as follows:

Format N Triples:

  • Instance Types: List of URLs and their types (DBpedia, Freebase etc)

E.g:

   <YOUR_LINK> <www.w3.org/1999/02/22-rdf-syntax-ns#type> <DBpedia:Type> .
  • Labels: List of URLs and Labels

E.g:

   <YOUR_LINK> <www.w3.org/2000/01/rdf-schema#label> "Label"@en .
  • Redirects: List of URLs and their redirect pages

E.g:

   <YOUR_LINK> <dbpedia.org/ontology/wikiPageRedirects> <YOUR_LINK> .

  • Disambiguations List of URLs and their disambiguations pages

    .

XML Dump:

  • Wiki dump - (like Wikipedia Dump).

After preparing these files with your own data, "just" follow the internationalization guide available in DBpedia Spotlight wiki to create the index with your own data.

All the best,

Sandro Athaide
  • 131
  • 4