Otimização do ranking de documentos em máquinas de busca na web a partir da mineração de dados sensível a contextos

• Ano de defesa: 2009
• Autor(a) principal: Joao Guilherme Rodrigues Gallo
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Contextos; Recuperação da Informação; Mineração de Dados; Sessões de Busca; Recuperação da informação; Computação; Mineração de dados (Computação)
• Link de acesso: http://hdl.handle.net/1843/ESBF-A8XQ6Q

This work presents a new approach to rank documents in web search engines based on context sensitive data mining. It's novelty lies specially on the use of a set of previously used strategies which have not been put together yet: First, the correlation patterns among the terms are processed efficiently. Second, a data mining technique called association rules is used for creating semantic correlation of the terms. Third, the identification of the concept searched by the user is done based on terms submitted on a retrieval session. Finally, all the process is done without the explicit demand of extra information from the user. Experimental results show that our approach increases the average precision of the search results on the evaluated collections for all kinds of searches without a substantial increase of the process computational. Our results suggest that our model presents considerable gains for generic Web text collections