Data fusion method for digital gazetteer

Abstract

Nowadays, digital gazetteers are proliferating in number and applied broadly in the location-based services. They are distributed in departments of government and various vendors, and repetitive efforts have been made to establish them which cost a lot of time and money. Frequently, to retrieve all-around information about a place, users have to consult several gazetteers and combine the results. In addition, we often develop a new gazetteer by integrating two or more sources of gazetteer data. So the fundamental challenge with digital gazetteers is how to fuse data from multi-sources to achieve the single place/single entry relationship. In this paper, we focus on how to merge the descriptions of the common places from different sources of digital gazetteers. A fusion method is put forward to judge identity of geographical names and combine their descriptions, which is based on linguistics, ontology and spatial relations theory and starting from geographical names meanings with their name strings, place type and spatial relationships taken into account. Experiments show that the method has enhanced efficiency and accuracy of integration of geographical names data automatically, and improve the batch update of geographical names.