Improving Genetic Information Extraction with Proximity Filtering
Author Information
Author(s): Masseroli Marco, Kilicoglu Halil, Lang François-Michel, Rindflesch Thomas C
Primary Institution: Politecnico di Milano
Hypothesis
Can proximity filtering enhance the precision of extracted genetic relations from natural language processing?
Conclusion
The proposed proximity filtering method significantly improves the precision of extracted genetic relations, making the information more reliable for further biomedical research.
Supporting Evidence
- Precision increased from 41.95% to 70.75% when filtering for arguments at a distance of 1 phrase from the predicate.
- The filtering method was evaluated on 2,042 relations extracted from Medline citations.
- The study demonstrated that proximity to the predicate positively affects the accuracy of extracted relations.
Takeaway
This study shows that when we look closely at how words are related in sentences, we can find more accurate information about genes and diseases.
Methodology
The study used a natural language processing system called SemGen to extract genetic relations and applied a filtering technique based on the distance between arguments and predicates.
Limitations
The filtering method may not be as effective for relations identified through co-occurrence processing compared to those from natural language processing.
Statistical Information
P-Value
p<0.05
Confidence Interval
95% confidence intervals 37.17% to 46.73%
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website