Classifying disease outbreak reports using n-grams and semantic features

This paper explores the benefits of using n-grams and semantic features for the classification of disease outbreak reports, in the context of a text mining system - BioCaster - that identifies and tracks emerging infectious disease outbreaks from online news. We show that a combination of bag-of-words features, n-grams and semantic features, in conjunction with feature selection, improves classification accuracy at a statistically significant level when compared to previous work. A novel feature of the work reported in this paper is the use of a semantic tagger - the USAS tagger - to generate features.