摘要：This paper is involved with extracting spatial information from text. We seek to geo-reference all spatial entities mentioned in a piece of text. The focus of this paper is to investigate the contribution of spatial and ontological reasoning to spatial interpretation of text. A preliminary study considering descriptions of cities and geographical regions from English Wikipedia suggests that spatial and ontological reasoning can be more effective to resolve ambiguities in text than a classical text understanding pipeline relying on parsing.
关键词：spatial information extraction; geo-referencing; spatial reasoning
