A neural network approach to quality control charts

Abstract

In this paper Quality Control Charts without memory are compared to neural networks trained with the Backpropagation algorithm. Neural networks are used to decide whether a process is under statistical control or out of control. As only the last sample is used to decide upon the state of the production process, a comparison to Shewhart-control charts leads automatically to a comparison between statistical tests and neural networks. By using a combined control chart to control the process mean and the process variability, the kind of classifications of the kind of change is considered explicitly. Finally neural networks are used to classify the kind of change occurred, considering only the last sample.