Date of Award
2006
Degree Type
Thesis
Degree Name
Master of Science
Program
Computer Science
Supervisor
Robert Mercer
Abstract
Knowledge in the field of biomedicine is booming. In order to effectively stay on the frontiers of their fields, researchers need automated tools to manage the ever increasing quantity of scholarly literature. As an aid to developing these tools, we suggest that the rhetoric found in the biomedical genre needs to be studied. This study is enabled by the availability of full text repositories such as BioMed Central and PubMed Central. To this end, this thesis proposes a model of rhetoric that captures the descriptive purpose of each sentence within a biomedical research article. This model proposes five Rhetorical Zones: Domain Knowledge, Experimental Procedure, Experimental Results, Meta-Experimental Results, and Meta-Experimental Procedure. This model was validated on a sample corpus of biomedical research articles with two machine-learning paradigms using a feature set of rhetorical cues extracted from the corpus. In addition, a system was developed to extract the feature set automatically.
Recommended Citation
Maynard, Trevor Alistair, "Distilling Rhetorical Information from Scholarly Biomedical Literature" (2006). Digitized Theses. 5041.
https://ir.lib.uwo.ca/digitizedtheses/5041