entity-fishing service updated. Spanish and Italian supported!
The NERD service entity-fishing automates the recognition and disambiguation of entities against Wikipedia and Wikidata using supervised machine learning. The tool is actively developed and is continuously evolving with new features and support to additional languages.
The team has finished the release of the version 0.0.3. Among the many improvements, the new version is now able to extract entities from texts not only in English, French and German, but also in Spanish and Italian.
The updated service is available at the address http://nerd.huma-num.fr/nerd.
Moreover, the new release contains:
- use of embeddings to improve the disambiguation
- improvements in the API messages/ error codes
- performances improvements
- fixes in results consistency
- bug fixes
For more information about the entity fishing service, you can read the following documentation: http://nerd.readthedocs.io/en/0.0.3 (especially in the upgrade guide: http://nerd.readthedocs.io/en/0.0.3/upgradeGuide.html).
For a general overview of the service and its implementation and integration on publishing platforms, we recommend you to look at the recording of the HIRMEOS webinar: http://www.hirmeos.eu/blog/webinars-presentations/