Output details
11 - Computer Science and Informatics
University of Sheffield
Disambiguation of biomedical text using diverse sources of information
<22> Determining the correct sense, in context, of words with multiple possible meanings is critical for applications such as automatic indexing and text mining. This paper [GoogleScholar: 23] reports work carried out within the EPSRC-funded BioWSD project on resolving lexical ambiguity in the biomedical domain and in collaboration with Martinez at Melbourne. It presents the first detailed study of how different information sources and their combinations contribute to disambiguation in biomedical text. It proposes a new information source that yielded a supervised learning model that outperformed the state-of-the-art. The BMC Bioinformatics has an impact factor of 3.02.