Challenges in Creating an Annotated Set of Geospatial Natural Language Descriptions (Short Paper)
Niloofar Aflaki, Shaun Russell, Kristin M. Stock · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · 2018
In order to extract and map location information from natural language descriptions, a first step is to identify different language elements within the descriptions. In this paper, we describe a method and discuss the challenges faced in creating an annotated set of geospatial natural language descriptions using manual tagging, with the purpose of supporting validation and machine learning approaches to annotation and text interpretation.