YODIE: Yet another Open Data Information Extraction system
1. YODIE in a Nutshell
YODIE performs named entity recognition and disambiguation against DBpedia across varied/noisy/short input text.
2. Try YODIE
Try this simple demo by pasting a text you'd like analysed (Because of limited computing resources, this demo is not always active).
3. How to Cite
Gorrell, Genevieve, Johann Petrak, and Kalina Bontcheva. "Using @Twitter Conventions to Improve #LOD-based Named Entity Disambiguation." In The Semantic Web. Latest Advances and New Domains, pp. 171-186. Springer International Publishing, 2015. [pdf]
4.1. English Tweets
The corpus introduced in the above publication is available here. It contains 794 tweets annotated with named entities disambiguated against DBpedia, and split into equally sized training and test portions. Tweets are in GATE FastInfoset format.
Users agree to abide by the Twitter TOS, and the annotated corpus is licensed by us under http://creativecommons.org/licenses/by-nc-sa/4.0/. Please cite the above paper if you use the corpus!
Research supported by the European Union under grant agreement No. 287863 TrendMiner, grant agreement No. 610829 Decarbonet, grant agreement No. 687847 COMRADES, and UK EPSRC grants Nos. EP/I004327/1.