1. Aplicación de la metodología no paramétrica Bootstrap en el control de calidad del proceso de envasado del café.
- Author
-
GONZALES CHAVESTA, CELSO, PACHECO OTÁROLA, VERÓNICA, and BALLENA, MANUEL
- Subjects
- *
STATISTICAL bootstrapping , *QUALITY control charts , *PACKAGING statistics , *COFFEE industry , *DATA distribution , *COFFEE - Abstract
The objective of the research is to show the efficiency of the Bootstrap methodology in the construction of limits of the control diagrams of the mean () - Range (R), its use in the control charts of Wu and Wang (1996) and to compare the results of this methodology with the classical form. It was found that the control limits for the R chart with the Bootstrap normal interval for 1000 (LCL = 0; UCL = 13.39) and 10000 (LCL = 0; UCL = 13.49), present a slight increase in the weight variation; while, with the Bootstrap percentile interval for 1000 (LCL = 0.1; UCL = 12.1) and 10000 (LCL = 0.1; UCL = 11.8), it presents a slight decrease in the variability of the weights respectively. Likewise, the control limits of the Bootstrap average with the Bootstrap normal interval for 1000 (LCL = 195.54; UCL = 207.42) and 10000 (LCL = 195.8; UCL = 207.2) which presents a slight stability in the average of the weights while than, with the Bootstrap percentile interval for 1000 (LCL = 196.6; UCL = 206.7) and 10000 (LCL = 196.6; UCL = 206.3). Therefore, the mean-range control diagram, when the assumption of normality is not fulfilled for small samples, has shown ineffectiveness in the detection and identification of special causes in a process. The research presents a non-parametric methodology that identifies the out-of-control signal (whether or not the assumption of normality), through the Bootstrap methodology in the construction of average-range control limits. [ABSTRACT FROM AUTHOR]
- Published
- 2020
- Full Text
- View/download PDF