• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 1
  • Tagged with
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

PragmaSUM: novos m?todos na utiliza??o de palavras-chave na sumariza??o autom?tica

Rocha, Valdir J?nior Cordeiro 05 December 2017 (has links)
Submitted by Jos? Henrique Henrique (jose.neves@ufvjm.edu.br) on 2018-05-03T18:35:26Z No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) valdir_junior_cordeiro_rocha.pdf: 3757934 bytes, checksum: 00a2e6ee18188436daa1415ec6a05021 (MD5) / Approved for entry into archive by Rodrigo Martins Cruz (rodrigo.cruz@ufvjm.edu.br) on 2018-05-04T16:22:37Z (GMT) No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) valdir_junior_cordeiro_rocha.pdf: 3757934 bytes, checksum: 00a2e6ee18188436daa1415ec6a05021 (MD5) / Made available in DSpace on 2018-05-04T16:22:37Z (GMT). No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) valdir_junior_cordeiro_rocha.pdf: 3757934 bytes, checksum: 00a2e6ee18188436daa1415ec6a05021 (MD5) Previous issue date: 2017 / Com a amplia??o do acesso ? internet e a cria??o de ferramentas que possibilitam pessoas a criarem conte?do, a informa??o dispon?vel cresce de forma acelerada. Textos sobre os mais diversos assuntos e autores s?o criados todos os dias. ? imposs?vel absorver a quantidade de informa??o dispon?vel, o que dificulta a escolha da mais adequada para determinado interesse ou p?blico. A sumariza??o autom?tica de textos, al?m de apresentar um texto de forma condensada, pode simplifica-lo, gerando uma alternativa para ganho de tempo e amplia??o do acesso a informa??o contida aos mais diferentes tipos de leitores. Os sumarizadores autom?ticos existentes atualmente na literatura n?o apresentam m?todos de personifica??o dos sum?rios para cada tipo de leitor, e consequentemente geram resultados pouco precisos. Este trabalho tem como objetivo utilizar o sumarizador autom?tico de textos PragmaSUM em textos educacionais com novas t?cnicas de sumariza??o utilizando palavras-chave. A utiliza??o de m?todos de personifica??o do sum?rio com palavras-chave visa aumentar a precis?o e melhorar o desempenho do PragmaSUM e seus sum?rios. Para isto, um corpus formado apenas por artigos cient?ficos da ?rea educacional foi criado para realiza??o de testes e compara??es entre diferentes sumarizadores e m?todos de sumariza??o. O desempenho dos sumarizadores foi medido pelas m?tricas Recall, Precision e F-Measure presentes na ferramenta ROUGE e validados com os testes estat?sticos ANOVA de Friedman e Coeficiente de Concord?ncia de Kendall. Os resultados obtidos apontam uma melhora no desempenho com a utiliza??o de palavras-chave na sumariza??o com o PragmaSUM, indicando a import?ncia na escolha adequada destas palavras-chave para classifica??o do conte?do do texto fonte. / Disserta??o (Mestrado Profissional) ? Programa de P?s-Gradua??o em Educa??o, Universidade Federal dos Vales do Jequitinhonha e Mucuri, 2017. / By expanding access to the internet and creating tools that enable people to create content, available information grows rapidly. Texts on the most diverse subjects and authors are created every day. It is impossible to absorb the amount of information available, which makes it difficult to choose the most appropriate for a particular interest or public. Automatic text summarization, as well as presenting a condensed text, can simplify it, generating an alternative to gain time and increase the access to information contained to the most different types of readers. The automatic summarizers that currently exist in the literature do not present methods of personification of the summaries for each type of reader, and consequently generate results inaccurate. This work aims to use the PragmaSUM automatic text summarizer in educational texts with new summarization techniques using keywords. Using summary keywords impersonation methods is intended to increase accuracy and improve the performance of PragmaSUM and its summaries. For this, a corpus formed only by scientific articles of the educational area was created to carry out tests and comparisons between different summarizers and summarization methods. The performance of the summarizers was measured by the Recall, Precision and F-Measure metrics present in the ROUGE tool and validated with the Friedman ANOVA statistical tests and Kendall's coefficient of agreement. The results obtained indicate an improvement in the performance with the use of keywords in the summarization with PragmaSUM, pointing out importance in the appropriate choice of these keywords for classification of the content of the source text.

Page generated in 0.0267 seconds