Inproceedings3765
Annotating Domain-Specific Texts with Babelfy: A Case Study
Annotating Domain-Specific Texts with Babelfy: A Case Study
Published: 2018
Buchtitel: Proceedings of the 1st International Workshop on EntitY REtrieval (EYRE 2018)
Verlag: n/a
Referierte Veröffentlichung
BibTeX
Kurzfassung
Freely available large knowledge graphs, such as DBpedia, Wikidata, and YAGO, generally provide a very solid representation of general knowledge, making them a good basis for text annotation. However, when it comes to annotating domain-specific text documents, these knowledge graphs need to be used with care. Moreover, publications describing real-world use cases of entity linking based on such knowledge graphs are surprisingly rare. In this paper, we describe the use case of annotating customer feedback texts written in German based on Babelfy as the text annotation service. We perform a manual evaluation of the annotations and show that Babelfy annotates around 85% of all annotations correctly. This makes Babelfy as a text annotation method and BabelNet as its knowledge graph a valid baseline for developing a custom knowledge graph and entity linking method, respectively.
Download: Media:Babelfy_EYRE2018.pdf
Weitere Informationen unter: Link
Information Retrieval, Semantische Annotation