Related Experiment Videos
Automatic lexicon acquisition for a medical cross-language information retrieval system
Kornél Markó1, Stefan Schulz, Udo Hahn
1Medical Informatics Department, Freiburg University Hospital, Germany. marko@coling.uni-freiburg.de
Studies in Health Technology and Informatics
|September 15, 2005
Summary
This study introduces an automated method for creating multilingual medical lexicons for Spanish and Swedish. This approach enhances medical cross-language information retrieval systems by leveraging existing resources and parallel texts.
Area of Science:
- Computational linguistics
- Medical informatics
- Natural language processing
Background:
- Medical cross-language information retrieval (CLIR) systems require robust multilingual medical lexicons.
- Existing methods for lexicon acquisition can be labor-intensive and time-consuming.
Purpose of the Study:
- To develop an automated method for acquiring multilingual medical lexicons for Spanish and Swedish.
- To enhance the performance of medical CLIR systems.
Main Methods:
- Utilizing seed lexicons and parallel corpora derived from the Unified Medical Language System (UMLS) Metathesaurus.
- Automatically generating Spanish and Swedish seed lexicons from existing Portuguese, German, and English resources.
- Iteratively validating lexical and semantic hypotheses using co-occurrence patterns in parallel corpora.
Main Results:
- Successful automated acquisition of multilingual medical lexicons for Spanish and Swedish.
- Demonstrated the feasibility of generating lexicons from diverse language sources.
- Validated the effectiveness of co-occurrence patterns for hypothesis validation.
Conclusions:
- The proposed automated method provides an efficient approach to building medical lexicons for CLIR.
- This technique can be extended to other languages and medical domains.
- Automated lexicon acquisition is crucial for advancing global medical information access.