http://lattes.cnpq.br/8567542948019912; FERNANDES, Ricardo Madeira.
Abstract:
Nowadays, search engines have increased their capability of indexing data available on the Web, making the task of providing relevant information to users even more difficult.
Advanced search options, such as date, language and geographical region filters, might
help to minimize this issue. However, there is still a need to develop search engines that
employ more sophisticated techniques than just keyword-based. The growth of social
networks and resource tagging turn the user as a key entity in this process, so that contextaware approaches are present in the recent research on Information Retrieval (IR), along with the well established ones, as the syntactic and semantic analysis of query terms. Nonetheless, the proposed solutions are, mainly, academic prototypes focused on only one approach. Hence, this work aims to develop a semantic and context-aware search engine, which combines techniques of IR literature with a new tag-based query expansion method. In order to validate the proposed ideas, the Geographic Information Retrieval System, called GeoSen, was extended with the provided features.