Serviços Personalizados
Journal
Artigo
Indicadores
- Citado por SciELO
- Acessos
Links relacionados
- Similares em SciELO
Compartilhar
Computación y Sistemas
versão On-line ISSN 2007-9737versão impressa ISSN 1405-5546
Resumo
LLANES, Raisa Socorro e ANDRES, Luisa Micó. Efficient use of Pivots for Approximate Search in Metric Spaces. Comp. y Sist. [online]. 2013, vol.17, n.4, pp.477-488. ISSN 2007-9737.
This work focuses on pivot-based fast nearest neighbor search algorithms that can work in any metric space. One of the objectives of these algorithms is to reduce the time consumed during search. Reducing time consumption of such algorithms usually consists in reducing the number of distances for computing, due to the high cost that they have in certain applications. We introduce a new version and improvements for a recently proposed algorithm, PiAESA, a variant of the AESA algorithm, used as baseline for performance measurement for over twenty years. The new version is simpler and allows better understanding of the algorithm and parameters used. Moreover, the efficiency is increased by defining an approximated version. Our empirical results with real and artificial databases confirm a consistent improvement in performance, when retrieving very high percentage of the correct answers (given by the exact algorithm).
Palavras-chave : Approximate search; metric spaces; near neighbor; distances.