We have done couple of researches here:
1- Focus locality extraction
2- Geotagging and Geoparsing
In every moment, several events are happening all over the world, and they are being reported by local or international news agencies. Automatic geolocation information extraction is vital to detect and analyze the impact of these events. We study extracting focus geolocations automatically from unstructured text-based news reports. The focus geolocation for a given news report is a location where an event mentioned in the report has happened.
This task is extracting geolocation from text and disambiguating the geolocation names. Geoparsers are Natural Language Processing (NLP) systems designed to analyze unstructured text in order to extract occurrences of location entities and place names and to resolve their ambiguities. Homographs and toponyms pose well-known problems in the process of geo-tagging news articles.
Here we compare different geoparsers performance with each other for English dataset.