Knowledge-based rules for the extraction of complex, fine-grained locative references from tweets

Abstract

The automatic analysis of user-generated text content from social media involves the challenge of extracting the locative references mentioned in microtexts, so that their geographic coordinates can be identified and the locations can be pinpointed on a map in geolocation systems. The goal of this article is to describe a knowledge-based model that captures a wide variety of locative references, ranging from geopolitical entities and natural landforms to points of interest and traffic ways, from English and Spanish tweets.