Automated placename resolution in text: geoparsing the Gazetteer for Scotland
Diss-1-research paper.doc (1018.Kb)
Diss-2-technical report.doc (1.390Mb)
MetadataShow full item record
The majority of textual information currently available on the Internet contains some sort of geographical information. Due to the unstructured nature of this information, compared to both the complexities of human languages and the ambiguity of placenames, the process of identifying them in a text is not a deterministic process. With the diffusion of Geographical Information Systems and underpinned by other disciplines like Informatics and Applied Linguistics, a number of commercial and academic projects have emerged, which focus on placename resolution, also known as geoparsing. This project will explore the complexities of geoparsing using the relational database of the Gazetteer for Scotland and, as a proof of feasibility, code will be produced and tested for this website, demonstrating the efficiency of the algorithms.