Automatic Annotation of Disposition Counts in News Articles.

Journal: Studies in health technology and informatics
Published Date:

Abstract

News media aggregate and report disposition counts during crises: how many people are affected, suspected affected, have died, and have recovered or been recovered; and they tend to do so in a timely and trustworthy manner. We present and evaluate a method for identifying these counts in unstructured natural language text, supporting downstream tasks such as automatic creation of epidemic curves.

Authors

  • Simon Rodier
    Digital Technologies, National Research Council Canada.
  • Dave Carter
    Digital Technologies, National Research Council Canada.