Processo de indução e ranqueamento de árvores de decisão sobre modelos OLAP

• Ano de defesa: 2010
• Autor(a) principal: Colares, Peterson Fernandes
• Orientador(a): Ruiz, Duncan Dubugras Alcoba
• 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: INFORMÁTICA; MINERAÇÃO DE DADOS (INFORMÁTICA); SISTEMA DE APOIO À DECISÃO (INFORMÁTICA); DATA WAREHOUSE; SISTEMAS DE INFORMAÇÃO
• Área do conhecimento CNPq: CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://tede2.pucrs.br/tede2/handle/tede/5163

Organizations acting on several markets have been using the benefits offered by the use of Data Mining - DM techniques as a complementary activity to their support systems to the strategic decision. However, to the great majority of the organizations, the deployment of a DM Project ends up not being feasible due to different factors, such as: Project duration, high costs and mainly by the uncertainty as to getting results that may effectively help the organization to improve their business processes. In this context, this paper presents a process based on the process of knowledge Discovery in Database - KDD which aims to identify opportunities to the application of DM techniques through the induction and ranking of decisions generated by the exploration of semi automatic Online Analytical Processing Models-OLAP. The built process uses stored information in a OLAP model prepared on the basis of used information by Customer Relationship Management - CRM and Business Intelligence - BI typically used by the organization to support strategic decision making. In relation to the information selected for this research, it has been carried out in a semi automatic way, a series of experiments using DM techniques which the results are collected and stored for later evaluation and ranking. The process was built and tested with a significant number of experiments and later evaluated by business experts in a large financial institution where this research was developed.