A double neural network approach for the identification and parameter estimation of control chart patterns
by Ahmed Shaban; Mohamed A. Shalaby
International Journal of Quality Engineering and Technology (IJQET), Vol. 3, No. 2, 2012

Abstract: The exhibited pattern on a control chart is classified as either natural or unnatural pattern. The presence of an unnatural pattern is evidence that a process is out of control. This paper devises neural networks as an intelligent tool to automate the identification of the different control chart patterns, and to accurately estimate their parameters. Two neural networks, named 'NN-1' and 'NN-2', are integrated together to perform the identification and the parameter estimation. The first stage 'NN-1' is developed to identify the existing pattern in the control data, and the second stage 'NN-2' is used to estimate the parameters of that pattern. NN-1 is developed to identify the five basic control chart patterns; namely: natural, upward shift, downward shift, upward trend, and downward trend. The probability of success in identifying the correct control charts pattern and its parameters is used to evaluate the performance of both NN-1and NN-2. Performance results of NN-1 and NN-2 are compared with other previous leading research work. Comparisons show that the proposed neural network approach yield better probability of success than the others.

Online publication date: Sat, 30-Aug-2014

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Quality Engineering and Technology (IJQET):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com