Ruben Verborgh’s data

Ruben’s data

Query Ruben’s data by triple pattern

Matches in Ruben’s data for { ?s ?p "Non-structured descriptive metadata provide additional benefits for end-user comprehension. However, their unstructured nature minimize their usefulness in an automated, digital context. This article explores the potential and the limits of Named Entity Recognition (NER) and Term Extraction (TE) in unstructured data searches in order to extract some meaningful concepts. These concepts allow us to benefit from improved retrieval and navigation, but they also play a very important role in digital humanities research. Using a case study to promote NER and TE experiments, based on the descriptive fields of the historical archives of Quebec City, the authors assess four third-party entity extractors. In an effort to address both NER and TE to assess named entities, they use a quantitative approach based on precision, recall and F-score calculated on the "gold standard corpus". A second more qualitative approach then leads us to consider the relevance of TE and to address the issue of multilingualism."@en }

Showing triples 1 to 1 of 1 with 100 triples per page.