Related Experiment Video
Updated: Aug 10, 2026
07:50
A Metadata Extraction Approach for Clinical Case Reports to Enable Advanced Understanding of Biomedical Concepts
Published on: September 20, 2018
Automatic extraction of acronym-meaning pairs from MEDLINE databases
J Pustejovsky1, J Castaño, B Cochran
1Laboratory for Linguistics and Computation at Brandeis University, Waltham, MA, USA.
Studies in Health Technology and Informatics
|October 18, 2001
Summary
This study introduces ACROMED, a system for extracting full forms of acronyms in biomedical texts. It achieves high precision and recall, outperforming other systems for specialized scientific language.
Area of Science:
- Biomedical Informatics
- Natural Language Processing
- Information Extraction
Background:
- Acronyms are prevalent in biomedical literature, posing challenges for automated information extraction.
- Existing methods for acronym expansion often struggle with the complex linguistic structures of scientific texts.
Purpose of the Study:
- To present ACROMED, a novel system for identifying and expanding acronyms within biomedical texts.
- To evaluate two distinct strategies for acronym long-form retrieval, tailored for the biomedical domain.
Main Methods:
- Developed ACROMED, integrating shallow parsing into acronym recognition.
- Tuned extraction strategies to accommodate the intricate phrase structures of the biomedical lexicon.
- Tested system performance on diverse biomedical text datasets.
Main Results:
- Achieved 72% recall and 97% precision in identifying acronym long forms.
- Demonstrated superior performance compared to general-purpose acronym extraction systems on biomedical texts.
Conclusions:
- ACROMED effectively addresses the challenge of acronym expansion in specialized biomedical literature.
- The system's performance highlights the benefits of domain-specific tuning and shallow parsing for information extraction.
