A neural network based model for abnormal pattern recognition of control charts

Abstract

In the past years, artificial neural networks were used for pattern recognition of control charts with an emphasis on recognizing specific abnormal patterns of control charts. This paper proposes an artificial neural network based model, which contains several back propagation networks, to both recognize the abnormal control chart patterns and estimate the parameters of abnormal patterns such as shift magnitude, trend slope, cycle amplitude and cycle length, so that the manufacturing process can be improved. Numerical results show that the proposed model also has a good recognition performance for mixed abnormal control chart patterns (e.g. a pattern with trend and cycle characteristics).