Buchtitel: Proceedings of the 23th International Conference on Theory and Practice of Digital Libraries (TPDL'19)
Referierte VeröffentlichungNote: https://dblp.org/rec/bibtex/conf/ercimdl/FarberS19
Citations have been classified based on their textual contexts w.r.t. their worthiness, function, polarity, and importance. To the best of our knowledge, so far citations have not automatically been classified by their grammatical role, that is, whether the citation (1) is grammatically integrated in the sentence, (2) is annotated directly after the occurrence of author names, (3) backs up a concept, (4) backs up a claim, or (5) is not appropriate because the context is incomplete or noisy.We argue that determining such classes for citation contexts is useful for a variety of tasks, such as improved citation recommendation and scientific impact quantification. In this paper, we propose this classification scheme, as well as a machine-learning-based approach to determine the classes automatically. Our evaluation reveals that the classification performance varies significantly between the citation types.
Weitere Informationen unter: Link
DOI Link: 10.1007/978-3-030-30760-8_38