Structured Wikinews
NLP processed version of English and Spanish Wikinews articles. We used a dump of 18,862 English and 7,603 Spanish articles. The data contains links to DBpedia for entities and provenance information as to in what source the event was mentioned. The entities and events in the Wikinews articles are structured using the…
NLP processed version of English and Spanish Wikinews articles. We used a dump of 18,862 English and 7,603 Spanish articles. The data contains links to DBpedia for entities and provenance information as to in what source the event was mentioned. The entities and events in the Wikinews articles are structured using the Simple Event Model (SEM) and the grounded annotation framework (GAF).
The English dataset contains:
- 811,885 Events
- 11,610 Locations
- 103,496 Actors
The Spanish dataset contains:
- 158,757
- 9,047 Locations
- 144,793 Actors
For each dataset, the original sources are available as a single cleaned xml file, the files containing the linguistic analyses as processed by a state-of-the-art document-based NLP pipeline and the aggregated information containing events and actors mentioned across different document (encoded as RDF/TRiG).
For more info and a list of files see: http://semanticweb.cs.vu.nl/structured-wikinews/
This dataset was created in NewsReader project (ICT-316404), funded by the European Union's 7th Framework Programme.
- cc-by-sa