Domain-specific language models and lexicons for tagging

• Published in: Journal of biomedical informatics Vol. 38; no. 6; pp. 422 - 430
• Main Authors: Coden, Anni R., Pakhomov, Serguei V., Ando, Rie K., Duffy, Patrick H., Chute, Christopher G.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.12.2005
• Subjects: Biomedical domain; Clinical information systems; Clinical report analysis; Corpus linguistics; Domain adaptation; Hidden Markov Model; Humans; Information Systems; Language; Markov Chains; Natural Language Processing; Part-of-speech tagging accuracy; Reproducibility of Results; Semantics; Statistical part-of-speech tagging; Unified Medical Language System; Part-of-speech tagging accuracy; Corpus linguistics; Statistical part-of-speech tagging; Domain adaptation; Biomedical domain; Clinical information systems; Hidden Markov Model; Clinical report analysis
• ISSN: 1532-0464; 1532-0480
• DOI: 10.1016/j.jbi.2005.02.009

Accurate and reliable part-of-speech tagging is useful for many Natural Language Processing (NLP) tasks that form the foundation of NLP-based approaches to information retrieval and data mining. In general, large annotated corpora are necessary to achieve desired part-of-speech tagger accuracy. We show that a large annotated general-English corpus is not sufficient for building a part-of-speech tagger model adequate for tagging documents from the medical domain. However, adding a quite small domain-specific corpus to a large general-English one boosts performance to over 92% accuracy from 87% in our studies. We also suggest a number of characteristics to quantify the similarities between a training corpus and the test data. These results give guidance for creating an appropriate corpus for building a part-of-speech tagger model that gives satisfactory accuracy results on a new domain at a relatively small cost.