Edited by Michel Aurnague and Dejan Stosic
[Human Cognitive Processing 66] 2019
► pp. 354–386
Geoparsing and geocoding places in a dynamic space context
The case of hiking descriptions
The backbone of the proposal in this chapter is an automatic parser and a formal encoder of information describing places, spatial and verbal relations in textual documents in order to reconstruct and map the textually described itinerary. These tools allow us to show how to combine the information expressed in French texts, referring to places, spatial actions associated with them, and data found in external geographical resources to build a geocoded representation of an itinerary. Our approach focuses on the automatic reconstruction of routes and transcribes them in their geographical setting, identifying locations and routes by interpreting spatial information in a dynamic space context.