Extracting Directional Relationships in Biomedical Literature
Author Information
Author(s): Giles Cory B, Wren Jonathan D
Primary Institution: Oklahoma Medical Research Foundation
Hypothesis
Can automated methods accurately extract directional relationships from large-scale biomedical literature?
Conclusion
The study demonstrates that while thesaurus-based directional relation extraction can achieve reasonable accuracy, it is susceptible to false positives due to noun modifiers.
Supporting Evidence
- The SVM classifier achieved 82% precision and 94.8% recall on gold standard corpora.
- Directional relationships were often extracted with ambiguity.
- Contextual factors influenced the interpretation of relationships.
Takeaway
The researchers created a computer program to find out how different biological things affect each other by reading a lot of scientific papers. They found that sometimes the results can be confusing.
Methodology
The study used a support vector machine (SVM) to classify dependency paths for extracting relationships from biomedical text.
Potential Biases
There is a risk of bias due to the reliance on noun modifiers and the context in which relationships are reported.
Limitations
The method is prone to false positives and may struggle with context-dependent relationships.
Statistical Information
P-Value
0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website