Predicting Domain Specific Entities with Limited Background Knowledge
Document Type
Paper
Abstract
This paper proposes a framework for automatic recognition of domain-specific entities from text, given limited background knowledge, e.g. in form of an ontology. The algorithm exploits several lightweight natural language processing techniques, such as tokenization and stemming, as well as statistical techniques, such as singular value decomposition (SVD) to suggest domain relatedness of unknown entities.
Publication Info
2006.
© Thomas, C., & Sheth, A. P., 2006
APA Citation
Thomas, C., & Sheth, A. P. (2006). Predicting Domain Specific Entities with Limited Background Knowledge.
https://corescholar.libraries.wright.edu/knoesis/37