Construction of large geographical database by merging Wikipedia's Geo-entities and GeoNames

Abstract

It is common to use on-line geographical databases for handling geographic information. For example, GeoNames is a free database for geographic information and Wikipedia is an encyclopedia that contains information of many geographic entities. However, since there is not enough linked information among these databases, it is not easy to combine the information stored in the both database. In this paper, we propose a method to find the corresponding entry between Wikipedia and GeoNames for constructing large Geographic database. In this approach, we made set of rules that represents the relationship between Wikipedia category information and feature class of GeoNames. We also propose a method to find out the corresponding Wikipedia's Geo-entities and GeoNames by using information of class correspondence, name, and country as constraints for matching.