1Department of Medical Informatics, Columbia University, New York, NY, USA.
You might also read
Articles linked to this work by shared authors, journal, and citation graph.
Choosing the right data representation is crucial for inductive learning algorithms classifying medical text. Explicitly capturing status information significantly improves classification performance, highlighting the impact of subtle data representation differences.
Area of Science:
Background:
Purpose of the Study:
Main Methods:
Main Results:
Conclusions: