Using parallel pivot vs. clustering-based techniques for web engines

Web Engines are a useful tool for searching information in the Web. But a great part of this information is non-textual and for that case a metric space is used. A metric space is a set where a notion of distance (called a metric) between elements of the set is defined. In this paper we present a...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Gil Costa, Graciela Verónica, Printista, Alicia Marcela
Formato: Objeto de conferencia
Lenguaje:Inglés
Publicado: 2007
Materias:
Ciencias Informáticas
Informática
Clustering
Search process
Parallel algorithms
metric spaces
parallel search
distance computations
Acceso en línea:http://sedici.unlp.edu.ar/handle/10915/22974
Aporte de:
SEDICI (UNLP) de Universidad Nacional de La Plata
id I19-R120-10915-22974
record_format dspace
institution Universidad Nacional de La Plata
institution_str I-19
repository_str R-120
collection SEDICI (UNLP)
language Inglés
topic Ciencias Informáticas
Informática
Clustering
Search process
Parallel algorithms
metric spaces
parallel search
distance computations
spellingShingle Ciencias Informáticas
Informática
Clustering
Search process
Parallel algorithms
metric spaces
parallel search
distance computations
Gil Costa, Graciela Verónica
Printista, Alicia Marcela
Using parallel pivot vs. clustering-based techniques for web engines
topic_facet Ciencias Informáticas
Informática
Clustering
Search process
Parallel algorithms
metric spaces
parallel search
distance computations
description Web Engines are a useful tool for searching information in the Web. But a great part of this information is non-textual and for that case a metric space is used. A metric space is a set where a notion of distance (called a metric) between elements of the set is defined. In this paper we present an efficient parallelization of a pivot-based method devised for this purpose which is called the Sparse Spatial Selection (SSS) strategy and we compare it with a clustering-based method, a parallel implementation of the Spatial Approximation Tree (SAT). We show that SAT compares favourably against the pivot data structures SSS. The experimental results were obtained on a highperformance cluster and using several metric spaces, that shows load balance parallel strategies for the SAT. The implementations are built upon the BSP parallel computing model, which shows efficient performance for this application domain and allows a precise evaluation of algorithms.
format Objeto de conferencia
Objeto de conferencia
author Gil Costa, Graciela Verónica
Printista, Alicia Marcela
author_facet Gil Costa, Graciela Verónica
Printista, Alicia Marcela
author_sort Gil Costa, Graciela Verónica
title Using parallel pivot vs. clustering-based techniques for web engines
title_short Using parallel pivot vs. clustering-based techniques for web engines
title_full Using parallel pivot vs. clustering-based techniques for web engines
title_fullStr Using parallel pivot vs. clustering-based techniques for web engines
title_full_unstemmed Using parallel pivot vs. clustering-based techniques for web engines
title_sort using parallel pivot vs. clustering-based techniques for web engines
publishDate 2007
url http://sedici.unlp.edu.ar/handle/10915/22974
work_keys_str_mv AT gilcostagracielaveronica usingparallelpivotvsclusteringbasedtechniquesforwebengines
AT printistaaliciamarcela usingparallelpivotvsclusteringbasedtechniquesforwebengines
bdutipo_str Repositorios
_version_ 1764820467953172482

Ejemplares similares