62013-01-23 18:59:48MarcinZajac52013-01-23 18:54:34MarcinZajac42013-01-23 18:48:14MarcinZajac32013-01-23 17:10:39MichalLenart22013-01-23 16:52:45MaciejOgrodniczuk12013-01-23 16:51:23MaciejOgrodniczuk

The DBPediaExtender is an information extraction system that extends an existing ontology of geographical entities by extracting information from text. The system uses distant supervision learning – training data is constructed based on matches between values from a knowledge base (DBPedia) and Wikipedia articles. The system was run on the Polish versions of DBPedia and Wikipedia and extracted more than 44 thousand RDF triples expressing relations between geographic entities from Polish Wikipedia.

Both the DBPediaExtender source code and generated triples are available under the GPLv3 licence. Author: Marcin Zając <mrcnzajac AT SPAMFREE gmail DOT com> License: GPLv3

A package containing both the source code and generated triples is available here. The package consists of: DBPediaExtender source code, user manual in English, 44174 RDF triples generated by the system. Project repository is available at .