Uncertainty Ontology for Veracity and Relevance
Erik Blasch, Carlos C. Insaurralde, Paulo Costa, Alta de Waal, J. P. de Villiers · 2019
The Uncertainty Representation and Reasoning Evaluation Framework (URREF) ontology addresses various semantic ambiguities associated with uncertainty. One of the main objectives of the URREF ontology is to define and articulate the criteria which enable the systematic reasoning about and evaluation of uncertainty representation and reasoning in information fusion systems. Towards assessing the merits of the URREF, information quality requires close attention for the measuring veracity and relevance. Veracity is measured from source sensitivity. Relevance is determined from precision and recall measures which are added to the URREF ontology. An example based on the Avionics Analytics Ontology (AAO) for Air Traffic Management with multiple Automatic Dependent Surveillance-Broadcast (ADS-B) radars is provided. The use or the precision-recall, F1 score, and sensitivity provide useful analysis for semantically important discussions on “relevance” as a complement of “recall” in information fusion uncertainty assessment.