Implementación de cartas de control en el paquete estadístico R para el monitoreo de procesos en media con datos autocorrelacionados

Statistical control charts are graps in statistical process control can monitoring quality characteristic(s) of an industrial process or service and are widely used today. To implement these assumes that the observations do not present an autocorrelation structure, but in practice this condition is violated frequently. The presence of autocorrelation has a serious impact on a substantial operation in the frequency of false alarms. Montgomery & Mastrangelo (1991) presents the construction of a control chart for autocorrelated data keeping in mind the structure of the EWMA statistic that is a good predictor one step ahead of the observations given. This control chart presents inconveniences when the autocorrelation is negative. In this paper we consider an alternative control chart which takes into account the data model through a time series ARMA family, for the construction of appropriate control limits from the negative correlation structure doesn’t cause problems.

• Autores: Gonzáles, Joaquín; Zambrano, Alex
• Fecha de publicación: 2011
• Institución: Universidad Santo Tomás
• Repositorio: Repositorio Institucional USTA
• Idioma: spa
• OAI Identifier: oai:repository.usta.edu.co:11634/39526
• Acceso en línea: https://revistas.usantotomas.edu.co/index.php/estadistica/article/view/16; http://hdl.handle.net/11634/39526
• Palabra clave: Cartas de control estadístico; autocorrelación; modelos ARMA; estadística EWMA; enseñanza y aprendizaje del control de calidad; paquete estadístico R
• License: http://purl.org/coar/access_right/c_abf2