Classification of News by Topic Using Location Data

In this work, we will consider news articles to determine geo-localization of their information and classify their topics on the basis of an available open data source: OpenStreetMap (OSM). We propose a knowledge-based conceptual and computational approach that disambiguates place names (i.e., geo-objects and regions) mentioned in news articles in terms of geographic coordinates. The geo-located news articles are analyzed to identify local topics: we found that the mentioned geo-objects are a good proxy to classify news topics.