Utilizando Evidência da wikipedia para relacionar textos a lugares

Abstract

Obtaining or approximating a geographic location for search results often motivates users to include place names and other geography-related terms in their queries. Previous work shows that queries that include geography-related terms correspond to a significant share of the users demand. Therefore, it is important to recognize the association of documents to places in order to adequately respond to such queries. This dissertation describes strategies for the geographic scope computation, using Wikipedia as an alternative source of direct and indirect geographic references. First we propose to perform a text classification task on geography-related classes, using textual evidence extracted from Wikipedia. We use terms that correspond to articles titles and the connections between articles in Wikipedias graph to establish a semantic network from which classification features are generated. Results of experiments using a news data-set, classified over Brazilian states, show that such terms constitute a valid evidence set for the geographic classification of documents, and demonstrate the potential of this technique for text classification. Another proposal describes a strategy for tagging documents with multiple place names, according to the geographic context of their textual content, using a topic indexing technique that considers Wikipedia articles as a controlled vocabulary. By identifying those topics in the text, we connect documents with the Wikipedia semantic network of articles, allowing us to perform operations on Wikipedias graph and find related places. We present an experimental evaluation on documents tagged as Brazilian states, demonstrating the feasibility of our proposal and opening the way to further research on geotagging based on semantic networks. Our results demonstrates the feasibility of using Wikipedia as an alternative source of geographical references. The method\\\'s main advantage is the use of free, up-to-date and wide knowledge and information from the digital encyclopedia. Finally, the Wikipedia introduction to the geographic text analysis can be faced as both, an alternative and a extension to use of geographical dictionaries (i. e. gazetteers).