Modelo computacional de recuperação da informação para repositórios digitais utilizando ontologias
Data
2017-02-17
Autores
Coneglian, Caio Saraiva [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
A evolução das Tecnologias da Informação e Comunicação causaram um aumento exponencial na produção e disseminação de dados na Internet. Dentre estas informações, inclui-se a produção científica que vive hoje um momento de transição, em que os documentos deixaram de ser apenas publicados em livros e revistas impressas e passaram a se espalhar pela rede. A partir disto, surgiu uma nova tecnologia chamada de repositórios digitais, em que são armazenados documentos em diversos formatos. Junto com o surgimento dos repositórios digitais, cresceu o desafio da recuperação destes documentos de maneira eficiente, ou seja, como a máquina poderá compreender o que o usuário procura, para fornecer os documentos que este usuário necessita. Neste âmbito, a Web Semântica surgiu visando possibilitar com que os computadores consigam compreender o contexto em que as informações criadas pelos usuários se encontram, tendo em suas ferramentas a base para tornar tal propósito real. No contexto dos repositórios digitais, esta pesquisa tem como objetivo aprimorar o processo de recuperação de informação nesses ambientes informacionais por meio da utilização do conceito de representações semânticas no uso de ontologias de domínio, que permita uma maior aderência na intersecção entre os itens bibliográficos e as necessidades informacionais dos usuários. Para atingir tais objetivos, utilizou-se uma metodologia de natureza quantitativa, em que se criou um modelo utilizando conceitos e tecnologias da Web Semântica para contextualizar o domínio da busca realizada pelo usuário. Como resultados, verificou-se que as relações das ontologias podem ser extraídas com eficiência por meio de um motor de geração de consultas SPARQL, que consegue localizar um termo na ontologia, bem como extrair as relações desse termo. Além disso, foram traçadas as ações que as propriedades do OWL devem possuir, no âmbito da recuperação da informação, para que assim possam ser identificadas com maior precisão as relações que um termo de busca possui frente a uma ontologia, permitindo a geração de uma nova expressão de busca, contendo um maior número de argumentos. Outro resultado obtido, diz respeito a interoperabilidade em repositórios digitais, que possibilitou identificar a integração e a recuperação dos metadados dos documentos dos repositórios digitais e a ferramenta tratando das questões semânticas. O trabalho propôs a interatividade na escolha das fontes informacionais, em que o usuário escolhe os repositórios em que seria realizada a busca, bem como cadastrar um repositório, caso este não tenha sido utilizado anteriormente. Conclui-se que a inserção de semântica em processos de recuperação de informação pode ocorrer por meio do modelo proposto, que se baseia essencialmente nas tecnologias e nos conceitos da Web Semântica, especialmente as ontologias, como um artefato capaz de explicitar o contexto em que os termos se encontram.
The evolution of Information and Communication Technologies has caused an exponential increase in the production and dissemination of data on the Internet. Among these data, we include the scientific production that is now in a moment of transition, in which the documents are no longer only published in printed books and magazines, and are now spread throughout the network. From this, a new technology has emerged called digital repositories, in which documents are stored in various formats. Along with the emergence of digital repositories, the challenge of recovering these documents has grown in an efficient way, i.e., how the machine can understand what the user is looking for, to provide the documents that this user needs. In this context, the Semantic Web came about in order to enable computers to understand the context in which the information created by the users meet, having in their tools the basis to make such a real purpose. In the context of Digital Repositories, this research aims to improve the Information Retrieval process in these informational environments through the use of the concept of semantic representations in the use of domain ontologies, which allows greater adherence at the intersection between bibliographic items and Information needs of users. In order to achieve these objectives, a quantitative methodology was used, in which a digital repositories interoperability model was created, using Semantic Web concepts and technologies to contextualize the search domain performed by the user. As results, it was verified that the relationships of the ontologies can be extracted efficiently by means of a SPARQL query engine, that is able to locate a term in the ontology, as well as to extract the relations of this term. In addition, we have outlined the actions that OWL properties must possess, in the context of Information Retrieval, so that the relationships that a search term has against an ontology can be identified more precisely, allowing the generation of a new Search expression, containing a greater number of arguments. Another result obtained concerns interoperability in digital repositories, which made it possible to identify the integration of the retrieval of the metadata of the digital repositories documents and the tool dealing with semantic issues. The work proposed the interactivity in the choice of informational sources, in which the user could choose the repositories in which the search would be carried out, as well as register a repository, if it had not previously been used. It is concluded that the insertion of semantics in Information Retrieval processes can occur through the proposed model, which is based essentially on the technologies and concepts of the Semantic Web, especially the ontologies, as an artifact capable of explaining the context in which the terms occur.
The evolution of Information and Communication Technologies has caused an exponential increase in the production and dissemination of data on the Internet. Among these data, we include the scientific production that is now in a moment of transition, in which the documents are no longer only published in printed books and magazines, and are now spread throughout the network. From this, a new technology has emerged called digital repositories, in which documents are stored in various formats. Along with the emergence of digital repositories, the challenge of recovering these documents has grown in an efficient way, i.e., how the machine can understand what the user is looking for, to provide the documents that this user needs. In this context, the Semantic Web came about in order to enable computers to understand the context in which the information created by the users meet, having in their tools the basis to make such a real purpose. In the context of Digital Repositories, this research aims to improve the Information Retrieval process in these informational environments through the use of the concept of semantic representations in the use of domain ontologies, which allows greater adherence at the intersection between bibliographic items and Information needs of users. In order to achieve these objectives, a quantitative methodology was used, in which a digital repositories interoperability model was created, using Semantic Web concepts and technologies to contextualize the search domain performed by the user. As results, it was verified that the relationships of the ontologies can be extracted efficiently by means of a SPARQL query engine, that is able to locate a term in the ontology, as well as to extract the relations of this term. In addition, we have outlined the actions that OWL properties must possess, in the context of Information Retrieval, so that the relationships that a search term has against an ontology can be identified more precisely, allowing the generation of a new Search expression, containing a greater number of arguments. Another result obtained concerns interoperability in digital repositories, which made it possible to identify the integration of the retrieval of the metadata of the digital repositories documents and the tool dealing with semantic issues. The work proposed the interactivity in the choice of informational sources, in which the user could choose the repositories in which the search would be carried out, as well as register a repository, if it had not previously been used. It is concluded that the insertion of semantics in Information Retrieval processes can occur through the proposed model, which is based essentially on the technologies and concepts of the Semantic Web, especially the ontologies, as an artifact capable of explaining the context in which the terms occur.
Descrição
Palavras-chave
Web semântica, Ontologias, Repositórios digitais, Recuperação da informação