A Field Based Representation for Vague Areas Defined by Spatial Prepositions

Natural language is one of the primary means of communicating spatial information, but existing geographic information retrieval (GIR) facilities are weak in this respect. One of the major challenges in automated interpetation of spatial natural language is how to model the regions implied by spatial expressions. This paper presents a field-based model for representing the vague regions defined by spatial language, including a method for defining field values from existing spatial language data sources. To interface this new vague field model with existing GIR systems and methods, an algorithm for extracting crisp boundaries from the field representation is also presented.