GeoXTag: Relative spatial information extraction and tagging of unstructured text
Syed M.A., Arsevska E., Roche M., Teisseire M.. 2022. In : Parseliunas E. (ed.), Mansourian A. (ed.), Partsinevelos P. (ed.), Suziedelyte-Visockiene J. (ed.). 25th AGILE Conference on Geographic Information Science “Artificial Intelligence in the service of Geospatial Technologies”. Göttingen : Copernicus Publications, p. 1-10. (AGILE: GIScience Series, 3). AGILE Conference on Geographic Information Science (AGILE 2022), 2022-06-14/2022-06-17, Vilnius (Lituanie).
Spatial information has gained more attention in natural language processing tasks in different interdisciplinary domains. Moreover, the spatial information is available in two forms: Absolute Spatial Information (ASI) e.g., Paris, London, and Germany and Relative Spatial Information (RSI) e.g., south of Paris, north Madrid and 80 km from Rome. Therefore, it is challenging to extract RSI from textual data and compute its geotagging. This paper presents two strategies and the associated prototypes to address the following tasks: 1) extraction of relative spatial information from textual data and 2) geotagging of this relative spatial information. Experiments show promising results for RSI extraction and tagging.
Documents associés
Communication de congrès
Agents Cirad, auteurs de cette publication :
- Arsevska Elena — Bios / UMR ASTRE
- Roche Mathieu — Es / UMR TETIS
- Syed Mehtab Alam — Es / UMR TETIS