Detecting Geographic Locations from Web Resources

by Wang et al. (GIR 2005)

The articles of the author distinguish between three different types of geographic locations

• the provider location (= source location),
• the content location (=target location), and
• the serving location which refers to the geographic scope that a Web resource can reach.

The authors include gazetteers which consider (i) geographic names, (ii) zip codes, and (iii) phone numbers and evaluate their approach by computing precision, recall, and the F1 measure.

Comments: Based on this article and work by Amitay we can identify the following steps which are required to detect geographic locations:

• geoparsing -- identification of geographic references in text,
• geocoding or grounding -- disambiguation of place names, and
• selection of the geographic focus -- assign the most relevant location to that document [amitay]