Modelo de mistura padrão de longa duração com censura uniforme-exponencial

• Ano de defesa: 2010
• Autor(a) principal: Chaves, Josenildo de Souza
• Orientador(a): Rodrigues, Josemar
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos
• Programa de Pós-Graduação: Programa de Pós-Graduação em Estatística - PPGEs
• País: BR
• Palavras-chave em Português: Análise de sobrevivência; Mistura de distribuições; Inferência bayesiana; Simulação estocástica
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::PROBABILIDADE E ESTATISTICA::ESTATISTICA
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/4482

In survival data analysis it is common the occurrence of a large number of individuals to the right. This fact can indicate that, in a fraction of the individuals the event of interest will never happen, in other words, a fraction of individuals of the population is cured or immune. This case is not usually taken into account by the usual survival theory that, in general, considers that the individuals at risk will not achieve cure during the follow-up period. Therefore, the survival models with cure fraction, or long-term survival models, have received a lot of attention in recent years. We consider the exponential distribution for the survival time of individuals at risk and the uniform-exponential distribution for the censoring time. In many situations, it is evident that the censoring mechanism is informative. Lagakos & Williams (1978) proposed a class of models where the acting of the censoring mechanism in the survival time is evaluated and Lagakos (1979) presented several situations in which the assumption of noninformative censoring is violated. The main purpose of this work is to verify the impact of informative uniform-exponential censoring in the survival data analysis under the standard mixture model.