Effects of non-normality on statistical and neural network based control chart

Loading...
Thumbnail Image
Date
2007
Authors
Khai Ping, Lim
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
In recent years, when quality becomes the priority in every field of manufacturing process, demand on quality control in non-normality is increasing accordingly. To achieve good quality control in non-normality, traditional control chart can no longer satisfy the user. In this dissertation, an artificial intelligent based control chart is introduced, developed and studied, which is Neural Network Control Chart. Neural network control chart is developed and analyzed using C++ programming in this dissertation. Then, the neural network control chart is compared with Shewhart control chart and memory control chart - EWMA in its performance in non-normality environment. The comparison criteria are average run length (ARL) performance during in-control stage, detection speed when out-of-control and type I errors when detecting the shift. The results in this dissertation show that neural network control chart performs better than the others in analyzing the non-normal data. It is good in low false signal during in-control stage, low rejection of in-control data at the same time quick detection when out-of-control, and consistent performance for different shapes of non-normal observations.
Description
Keywords
Non-normality on statistical , Network based control chart
Citation