Encontrando os locais de interesse com maior concentração de objetos relevantes para um conjunto de palavras-chave

Detalhes bibliográficos
Ano de defesa: 2019
Autor(a) principal: Andrade, Claudio Moisés Valiense lattes
Orientador(a): Rocha Junior, João Batista lattes
Banca de defesa: Não Informado pela instituição
Tipo de documento: Dissertação
Tipo de acesso: Acesso aberto
Idioma: por
Instituição de defesa: Universidade Estadual de Feira de Santana
Programa de Pós-Graduação: Mestrado em Computação Aplicada
Departamento: DEPARTAMENTO DE CIÊNCIAS EXATAS
País: Brasil
Palavras-chave em Português:
Área do conhecimento CNPq:
Link de acesso: http://tede2.uefs.br:8080/handle/tede/826
Resumo: Spatial data is increasingly present in our daily lives. We use various applications that use this data, such as Google Maps and Uber. There are a huge number of interesting questions that can be performed on these data. For example, a tourist maybe interested in hotels that have a lot of restaurants in its vicinity. This project proposes a new query type named Popularity Based Spatio-Textual Preference Query (PSTPQ), whose main contribution, that can select the best spatial objects taking into account the number of relevant spatio-textual objects, for a given set of query keywords, in its neighborhood. We present algorithms to process this query e ciently and evaluate the algorithms proposed in real datasets. Our experiments show that it is more e cient to use spatial indices (e.g. R-Tree) for distances less than 5 km in relation to textual indexes (e.g. Inverted File). In our experiments, the hybrid index processed the PSTPQ query with better performance. The PSTPQ query has as a di erential take into account the number of reference objects in the spatial neighborhood, in addition to selecting the reference objects from the textual description.