GeoST: Geographic, Thematic and Temporal Information Retrieval from Heterogeneous Web Data Sources

Abstract

AbstractGeographical data commonly available on the Web is often related to thematic and temporal data, and represented in heterogeneous data sources. This makes geographical information retrieval tasks non straightforward processes as the information available is often unstructured. The research presented in this paper introduces an ontology-driven approach whose objective is to facilitate retrieval of geographical information on the Web. Spatio- temporal queries are categorized in order to solve questions related to several complementary modeling dimensions: when, what andwhere. The retrieval strategy is based on the exploration of a spatio-temporal ontology, and implemented by a search engine. The whole approach is exemplified by a prototype development, and the results obtained are evaluated by a metrics based on recall and precision measures. The experimental findings show that the strategy is effective for the types of spatio-temporal queries used by our geographical retrieval approach.KeywordsGeographic and temporal information retrievalOntologyWeb information retrieval