Registro Completo |
Biblioteca(s): |
Embrapa Agricultura Digital. |
Data corrente: |
27/11/1997 |
Data da última atualização: |
19/12/2007 |
Autoria: |
GAMMOUDI, M. M.; AQUINO, D. C. |
Título: |
Formal method for document clustering based on their semantics and their use in information retrieval. |
Ano de publicação: |
1996 |
Fonte/Imprenta: |
In: SIMPÓSIO BRASILEIRO DE BANCO DE DADOS, 11., 1996, São Carlos, SP. Anais... São Carlos: USP-ICMSC, 1996. |
Páginas: |
p.396-410. |
Idioma: |
Inglês |
Notas: |
SBBD'96. Editado por Teresa Pires Vieira e Agma Juci Machado Traina. |
Conteúdo: |
Research works in Information Retrieval Systems (IRS) show that cluster mechanism of documents is very important to perform information access [22]. Some methods are proposed which are based on the concept of similarity between documents or between indexing terms. They are used in several IRS. However, they have some limitations such as the lack of mathematical justification of formulate used to compute similarities. In this paper, we introduce a formal method called Rectangular Decomposition of a Binary Relation (RDBR)\tfor simultaneous document and indexing term clustering. This method provides a combination of inverted file generation and clustering, which is an interesting alternative to be procedures currently in use. This method is based on two heçuriuristics in the objective to give an approximative solution for rectangle extraction from a binary relation which is an NP-complete problem [1, [1,9]. |
Palavras-Chave: |
Banco de dados. |
Thesaurus Nal: |
databases. |
Categoria do assunto: |
-- |
Marc: |
LEADER 01555naa a2200181 a 4500 001 1005402 005 2007-12-19 008 1996 bl uuuu u00u1 u #d 100 1 $aGAMMOUDI, M. M. 245 $aFormal method for document clustering based on their semantics and their use in information retrieval. 260 $c1996 300 $ap.396-410. 500 $aSBBD'96. Editado por Teresa Pires Vieira e Agma Juci Machado Traina. 520 $aResearch works in Information Retrieval Systems (IRS) show that cluster mechanism of documents is very important to perform information access [22]. Some methods are proposed which are based on the concept of similarity between documents or between indexing terms. They are used in several IRS. However, they have some limitations such as the lack of mathematical justification of formulate used to compute similarities. In this paper, we introduce a formal method called Rectangular Decomposition of a Binary Relation (RDBR)\tfor simultaneous document and indexing term clustering. This method provides a combination of inverted file generation and clustering, which is an interesting alternative to be procedures currently in use. This method is based on two heçuriuristics in the objective to give an approximative solution for rectangle extraction from a binary relation which is an NP-complete problem [1, [1,9]. 650 $adatabases 653 $aBanco de dados 700 1 $aAQUINO, D. C. 773 $tIn: SIMPÓSIO BRASILEIRO DE BANCO DE DADOS, 11., 1996, São Carlos, SP. Anais... São Carlos: USP-ICMSC, 1996.
Download
Esconder MarcMostrar Marc Completo |
Registro original: |
Embrapa Agricultura Digital (CNPTIA) |
|