Towards a representation of uncertain geospatial information in knowledge graphs

Abstract

This paper highlights the challenges of representing uncertain geospatial information in knowledge graphs. We propose to use Real Estate advertisements since professionals use a lot of vernacular and vague places in order to promote a house to their target audience. Then, we suggest to model local place names using fuzzy set theory. Finally, we discuss how to build a knowledge graph that represents extracted geospatial objects and their uncertainty.