Detalhes bibliográficos
Ano de defesa: |
2013 |
Autor(a) principal: |
Amaral, Daniela Oliveira Ferreira do
 |
Orientador(a): |
Vieira, Renata
 |
Banca de defesa: |
Não Informado pela instituição |
Tipo de documento: |
Dissertação
|
Tipo de acesso: |
Acesso aberto |
Idioma: |
por |
Instituição de defesa: |
Pontifícia Universidade Católica do Rio Grande do Sul
|
Programa de Pós-Graduação: |
Programa de Pós-Graduação em Ciência da Computação
|
Departamento: |
Faculdade de Informáca
|
País: |
BR
|
Palavras-chave em Português: |
|
Área do conhecimento CNPq: |
|
Link de acesso: |
http://tede2.pucrs.br/tede2/handle/tede/5246
|
Resumo: |
Many tasks in Natural Language Processing involves the provision of a large number of variables, which depend on each other. Structured prediction methods are essentially a combination of classification and modeling based on graphs. They combine the power of classification methods with the ability of this type of modeling to play compactly, multivariate data. The classification methods perform prediction using a large set of features as input. Conditional Random Fields (CRF) is a probabilistic method for predicting structured and has been widely applied in various areas such as natural language processing, including the Named Entity Recognition (NER), computer vision, and bioinformatics. Therefore, this dissertation proposes the application of CRF to NER for the Portuguese Language and to evaluate their performance based on the HAREM corpus. Finally, comparative tests of similar approaches were performed, illustrating the efficiency and competitiveness of the proposed system. |