Gráfico de Hotelling com esquemas especiais de amostragem para o monitoramento de processos bivariados autocorrelacionados

• Ano de defesa: 2015
• Autor(a) principal: Leoni, Roberto Campos [UNESP]
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Amostragem (Estatistica); Autocorrelação (Estatistica); Analise multivariada; Graficos aleatorios; Variaveis aleatorias; Autocorrelation (Statistics)
• Link de acesso: http://hdl.handle.net/11449/126617; http://www.athena.biblioteca.unesp.br/exlibris/bd/cathedra/05-08-2015/000839268.pdf

The most important of the assumptions made concerning control charts is that of independence of the quality characteristics observations. The violation of the independence assumption decreases the power of the control chart. In this thesis, is considered the T2 control chart for bivariate samples of size n with observations modeled by a first order vector autoregressive model - VAR (1). To counteract the undesired effect of the autocorrelation two sampling strategies are applied: (1) the systematic sampling strategy, where the samples are obtained by selecting an element of the production line and skipping s consecutive elements before selecting the next one; (2) the mixed sampling strategy where the samples elements are selected from the two consecutive rational subgroups. The sampling strategies always improves the chart's performance, except when only one variable is affected by the assignable cause and the observations of this variable are not autocorrelated. If only one element is skipped, the average run length (ARL) reduces in more than 30%, on average. If two elements are skipped, this number increases to 40%. With the mixed sample, the average reduction is 25% in the ARL