Medical drug reviews are increasingly commonplace on the web and have become
an important source of information for patients undergoing medical treatment. Patients will look to these reviews in order to understand the impact the drugs have
had on others who have experienced them. In short these reviews can be interpreted
as a body of arguments and counterarguments for/against the drug being reviewed.
One of the challenges of reading these reviews is drawing out the arguments easily
and forming a final opinion; this is due to the number of reviews and the variety of
arguments presented.
This thesis explores the use of computational models of argumentation in order
to extract structured argumentation data from the reviews and present them to the
user. In particular I propose a pipeline that performs argument extraction, argument
graph extraction and visualisation.