Towards a Semantic Lexicon for Biological Language Processing
Author Information
Author(s): Karin Verspoor
Primary Institution: Los Alamos National Laboratory
Hypothesis
Can the Unified Medical Language System (UMLS) be effectively used to construct a lexicon for processing biological texts?
Conclusion
The constructed lexicon covers over 77% of the most frequent tokens in the biological domain corpus, indicating its potential utility for biological text processing.
Supporting Evidence
- Over 77% of tokens in the domain corpus are found in the constructed lexicon.
- Only 3% of the unique tokens in the corpus are covered by the lexicon.
- The lexicon includes the main content-bearing terms of the domain.
Takeaway
This study created a dictionary to help computers understand biological texts better, and it found that most important words in the texts are included in this dictionary.
Methodology
The study analyzed the overlap between the UMLS Metathesaurus and the SPECIALIST lexicon to create a lexicon for biological language processing, using a corpus of full-text articles from the Journal of Biological Chemistry.
Limitations
The corpus used may not fully represent the entire domain of molecular biology, and the study did not assess the utility of the UMLS Semantic Network for biological language processing.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website