Creating a Medical Dictionary Using Word Alignment
Author Information
Author(s): Mikael Nyström, Magnus Merkel, Håkan Petersson, Hans Åhlfeldt
Primary Institution: Linköpings universitet
Hypothesis
Can different types of resources improve the quality of automatic word alignment for medical terminology systems?
Conclusion
More resources give better results in the automatic word alignment, but some resources only give small improvements.
Supporting Evidence
- The verified English-Swedish dictionary contains 24,000 term pairs in base forms.
- Manually created training resources give nearly as good results as a union of static and statistical resources.
- More resources and resource types generally lead to better results in automatic word alignment.
Takeaway
Using more resources helps create better translations between English and Swedish medical terms, but some resources don't help much.
Methodology
The study involved automatic word alignment of medical terminology systems using various types of resources, followed by manual verification of candidate term pairs.
Limitations
The study may not generalize to all medical terminology systems due to differences in structure and content.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website