IRank: Integral ranking of geographical information by semantic, geographic, and topological matching

Previous geographic information retrieval (GIR) works have used different criteria of a geographical nature to rank the documents retrieved from heterogeneous repositories. The most common approaches consider the characteristics and relationships that describe the geographical objects. However, these criteria process the documents in a separate way (only using their geometric or topologic aspects). In addition, they do not take into account the nature of geographic data (spatial semantics) in the weighting and ranking process which limits the assessment of document relevance. Nevertheless, the ranking can be improved by using approaches integrating the essence and nature of geographical space, i.e., (1) geographical attributes, (2) topological relationships, and (3) spatial semantics that are focused on conceptually describing a geographic object. This paper outlines iRank, a method that integrates these three aspects to rank a document. iRank evaluates documents using three sources of information: GeoOntologies, dictionaries, and topology files. The approach consists of three stages which define the geographical relevance between a query and a document. In the first stage, the relevance is computed by using concepts (GeoOntologies), the second stage uses geographic attributes (dictionaries), and in the last stage, the relevance is processed by considering spatial relationships (vector files). Thus, the major iRank advantage is integral ranking. The results received by the authors show a better ranking with these criteria than ones that use them separately.