Inproceedings: an article in a conference proceedings.
Distinguishing general concepts from individuals: an automatic coarse-grained classifier
Title of the conference
16th International conference on knowledge engineering and knowledge management knowledge patterns, Acitrezza, 2008. Poster and demo proceedings
Named entity recognizers are unable to distinguish if a term is a general concept as "scientist" or an individual as "Einstein". In this paper we explore the possibility to reach this goal combining two basic approaches: (i) Super Sense Tagging (SST) and (ii) YAGO. Thanks to these two powerful tools we could automatically create a corpus set in order to train the SuperSense Tagger. The general F1 is over 76% and the model is publicly available.
Last modification date