VASTE- Digicosme Paris Saclay Project







VASTE : Veracity Assesment in Spatio-TEmporal heterogeneous data 

An application on Web animal epidemiological surveillance

A digicosme Paris Saclay project (2018-2019)

Summary

VASTE project aims at assessing the veracity of epidemiological events by exploiting the knowledge and the data coming from different data sources of different origins and different quality levels: structured data derived from expert reports published by official agencies (e.g. OIE, FAO, OMS) and data produced by a process of text mining from unofficial sources (e.g. local newspapers, blogs). For this purpose, we plan to develop an approach that will combine data linking approaches and reasoning mechanisms from argumentation theory. Indeed, defining new data linking methods which while considering data quality indicators (e.g., freshness, reliability) will allow to determine the groups of events referring to the same disease, the same species, the same localisation and the same period, and thus will be able to indicate how true is an event. The argumentation reasoning will allow enforcing the obtained truthfulness, by reasoning on positive and negative expert arguments. To prove the effectiveness and the efficiency of the proposed approach, an experimental evaluation will be conducted on datasets that were already collected by TETIS Lab on the animal epidemiological surveillance domain.

Project Partners 

  • Telecom ParisTech
  • AgroParisTech & INRA
  • LRI, Université Paris Sud & CNRS
  • DAVID, Université Versailles Saint-Quentin- en-Y velines
  • TETIS, Montpellier

Main contact: Fatiha SAIS (sais [AT] lri [DOT] fr)