Detecting Web Rumours with a Multilingual Ontology-Supported Text Classification System

Timely surveillance of disease outbreak events of public health concern currently requires detailed and time consuming manual analysis by experts. Recently in addition to traditional information sources, the World Wide Web has offered a new modality in surveillance, but the massive collection of multilingual texts which must be processed in real time presents an enormous challenge.

Objective

In this paper we present a summary of the BioCaster system architecture for Web rumour surveillance, the rationale for the choices made in the system design and an empirical evaluation of topic classification accuracy for a gold-standard of English and Vietnamese news.