An improved features selection approach for control chart patterns recognition

<span lang="EN-US">Control chart patterns (CCPs) are an essential diagnostic tool for process monitoring using statistical control (SPC). CCPs widely used to improve production quality in many engineering applications. The principle is recognize the state of a process, either stable or deterioration unstable process. It significantly narrow set possible assignable causes by shortening quality. Machine learning techniques have been CCPs. Artificial neural networks with multilayer perceptron (ANN-MLP) one standard tools this purpose. This paper proposes improved features selection method select best as input representation recognition. results demonstrate that proposed approach can effectively even small mean shift less than 1.5 sigma. dimensional reduction was achieved employing Relief, correlation, and Fisher algorithms (RCF) feature classifier (RCF-ANN). study provides experimental result compares performance before after reduction.</span>