A georeferencing multistage method for locating geographic context in web search

Abstract

The geographic scope of Web pages is becoming an essential dimension of Web search, especially for mobile users. This paper shows a multistage method for assigning a geographic focus to Web pages (GeoReferencing) according to their textual contents. We suggest several heuristics for the disambiguation of toponyms and a scoring procedure for focus determination. Furthermore, we provide an experimental methodology for evaluating the accuracy. Finally, we obtained promising results of over 70% accuracy with a city-level resolution.

A survey on the geographic scope of textual documents

“…For instance, if many states of a country are mentioned in a document, a more adequate GS(D) may indicate the country itself as the document's scope. Some other terms are equivalent: geographic document footprint (Fu et al, 2005;Silva et al, 2006), geographic path (Vargas et al, 2012b), and geographic focus (Amitay et al, 2004;Chen et al, 2010;Zubizarreta et al, 2008).…”

“…The Geographic Scope Resolution (Alexopoulos and Ruiz, 2012;Andogah et al, 2012;Alexopoulos et al, 2013) problem consists in discovering places related to the contents of a textual document, disambiguating them if necessary, and using the resulting set of unique places to build the overall geographic scope. GSR is known under other names in the literature, with essentially the same definition, except for terminological differences: Place Name Assignment Problem (Zong et al, 2005;Amitay et al, 2004) or GeoReferencing (Gouvêa et al, 2008;Zubizarreta et al, 2008).…”

“…Checking the validity of candidate names as indicators of location requires using reference datasets, such as gazetteers, ontologies or proper databases, along with matching algorithms and heuristics to take care of disambiguation. Strategies used to solve this problem are the use of a controlled dictionary as support in the geoparsing step and heuristics to resolve the ambiguity between the references (Zubizarreta et al, 2008), the use of a NER procedure and a graph-ranking algorithm, based on PageRank (Page et al, 1999) for the geoparsing and reference resolution steps (Silva et al, 2006). Zong et al (2005) use a third party software to perform the geoparsing and a rule-based approach to disambiguate the toponyms using a tree structure to build the geographic scope instead of the graph structure used by Zubizarreta et al (2008) and Silva et al (2006).…”

“…Strategies used to solve this problem are the use of a controlled dictionary as support in the geoparsing step and heuristics to resolve the ambiguity between the references (Zubizarreta et al, 2008), the use of a NER procedure and a graph-ranking algorithm, based on PageRank (Page et al, 1999) for the geoparsing and reference resolution steps (Silva et al, 2006). Zong et al (2005) use a third party software to perform the geoparsing and a rule-based approach to disambiguate the toponyms using a tree structure to build the geographic scope instead of the graph structure used by Zubizarreta et al (2008) and Silva et al (2006). Vargas et al (2012a), Vargas et al (2012b) and Zhang et al (2012) also have multistage methods using a third part tool in the geoparsing step.…”

Identificação da cobertura espacial de documentos usando mineração de textos

“…3.2.4 Extração de informação considerando o foco geográfico (Zubizarreta et al, 2008) No trabalho de Zubizarreta et al (2008) é explorado um método para atribuir um foco geográfico às páginas Web considerando o contexto. Nesse trabalho foram sugeridas varias heurísticas para o processo de desambiguação por topônimos.…”