Métodos de simulação de dados geográcos sintéticos para basescondenciais

• Ano de defesa: 2018
• Autor(a) principal: Leticia Silva Nunes
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Dados sintéticos; Coordenadas geográcas; Condencialidade; Privacidade; Método de Monte Carlo; Estatistica; Privacidade Informática
• Link de acesso: http://hdl.handle.net/1843/BUOS-AZGMYK

This work presents statistical methods to preserve condentiality of datasets, in particular simulation of synthetic data. This approach is based on simulate sinthetic data that are generated from from specied probability distributions to reproduce the maximum possible relationships contained in the original data. Simulation of syntheticdata is attractive because it allows controlling the disclosure risk and utility. Futhermore, researchers will be able to use conventional statistical analysis tools. In addition, we extended the methodology for simulation of synthetic geographical coordinates for spatial data that is proposed in Paiva et al. (2014) by inserting continuous variables in theestimation of the model