Semantic Similarities between Locations based on Ontology
| Authors | |
|---|---|
| Year of publication | 2017 |
| Type | Article in Proceedings |
| Conference | Proceedings of the Eleventh Workshop on Recent Advances in Slavonic Natural Language Processing, RASLAN 2017 |
| MU Faculty or unit | |
| Citation | |
| web | |
| Field | Informatics |
| Keywords | toponym disambiguation; geonames; geographic text retrieval; ontology based geoname relations; toponym similarity |
| Description | Toponym disambiguation or location names resolution is a critical task in unstructured text, articles or documents. Our research explores how to link ambiguous locations mentioned in documents, news and articles with latitude/longitude coordinates. We designed an evaluation system for toponym disambiguation based on annotated GEOCLEF data. We implemented a node-based approach taking population into account and a geographic distance-based approach. We have proposed new approach based on edges between the pairs of toponyms in ontology, taking also population attribute into account. Our edge-based approach gave better results than population and distance-based only approaches. The results could be used in any information system dealing with texts containing geographic locations, such as news texts. |
| Related projects: |