Semantator: annotating clinical narratives with semantic web ontologies by Dezhao Song, Christopher G. Chute, and Cui Tao. (AMIA Summits Transl Sci Proc. 2012;2012:20-9. Epub 2012 Mar 19.)
Abstract:
To facilitate clinical research, clinical data needs to be stored in a machine processable and understandable way. Manual annotating clinical data is time consuming. Automatic approaches (e.g., Natural Language Processing systems) have been adopted to convert such data into structured formats; however, the quality of such automatically extracted data may not always be satisfying. In this paper, we propose Semantator, a semi-automatic tool for document annotation with Semantic Web ontologies. With a loaded free text document and an ontology, Semantator supports the creation/deletion of ontology instances for any document fragment, linking/disconnecting instances with the properties in the ontology, and also enables automatic annotation by connecting to the NCBO annotator and cTAKES. By representing annotations in Semantic Web standards, Semantator supports reasoning based upon the underlying semantics of the owl:disjointWith and owl:equivalentClass predicates. We present discussions based on user experiences of using Semantator.
If you are an AMIA member, see above for the paper. If not, see: Semantator: annotating clinical narratives with semantic web ontologies (PDF file). And the software/webpage: Semantator.
Software is a plugin for Protege 4.1 or higher.
Looking at the extensive screen shots at the website, which has good documentation, the first question I would ask a potential user is: “Are you comfortable with Protege?” If they aren’t I suspect you are going to invest a lot of time in teaching them ontologies and Protege. Just an FYI.
Complex authoring tools, particularly for newbies, seem like a non-starter to me. For example, why not have a standalone entity extractor (but don’t call it that, call it “I See You (ISY)) that uses a preloaded entity file to recognize entities in a text. Where there is uncertainty, those are displayed in a different color, with drop down options on possible other entities. User get to pick one from the list (no write in ballots). Performs a step towards getting clean data for a second round with another one-trick-pony tool. User contributes, we all benefit.
Which brings me to the common shortfall of annotation solutions: the requirement that the text to be annotated be in plain text.
There are lot of “text” documents but what of those in Word, PDF, Postscript, PPT, Excel, to say nothing of other formats?
The past will not disappear for want of a robust annotation solution.
Nor should it.