Web2DB : uma ferramenta para a construção de representações relacionais de sitios da web

• Ano de defesa: 2008
• Autor(a) principal: Marcelo Dias Correa
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: web; Representações relacionais; Recuperação de dados; Banco de dados relacionais; World Wide Web (Sistema de recuperação da informação); Recuperação de dados (Computação); Computação
• Link de acesso: http://hdl.handle.net/1843/RVMR-7KXND4

The increasing demand for valuable information to be used in the analysis and decision-making processes favors the development of tools and methods that automate the extraction and treatment of web data. The rise in Web's popularity has given place for an enormous quantity of documents widely spread over the Web. The centralization of the data is important because it reduces the efforts on retrieving the useful information from the vast repositories, allowing the efforts to focus more on the analysis and decision-making processes rather than lower-level data-handling techniques. In many cases the interest resides in an effective way to search for information rather than visiting unstructured web pages hoping to find the right data.The motivation for this work started from the need to create a process that would permit the collection of web pages containing the desired user data and the extraction of the data based on a relational representation previously configured. The resulting relational database could be analyzed and manipulated according to the needs of many applications. In this context it was designed Web2DB, a tool that, giving a model for a web site, permits the configuration and execution of page data collection and then the extraction of the data to a database. The user can customize the types of pages to be collected, the extraction interest data and the way which the database will be populated. The tool also permits the generation of views so the extracted data can the visualized in the most convenient way.The tool uses the example-based data extraction strategy. The user participation in the process-mapping phase is intended to aggregate value from the business model into the process. The following activities after the mapping phase are done automatically by the tool. This is a practical approach to the data extraction problem aiming the analysis of a vast diffuse data spread on web sites. The tool is suitable for relational representations of big web sites and, for being customizable, can be applied in most electronic sites that meet a list of requisites for the extraction.