Automatic Generation of Event Ontology from Social Network and Mobile Positioning Data

Abstract

The study of mobile positioning data makes it possible to detect whether an event has happened at a particular place during a given period. However, determining the nature and details of the event is a challenge, especially if the event is not widely known, as is the case for local events. We propose an approach to determining the nature of local events by generating an ontology in a completely automatic way from social network data and data on people's movements and by querying this generated ontology. This approach uses entity discovery techniques, filtering systems and information enrichment via Open Data, as well as a system for matching discovered entities and ontology elements. Evaluation via a survey allowed us to validate approximately that the information presented in the ontology is reliable, makes sense and answers our questions.