Applying Machine Learning Approaches to the Recognition of Variance Fault of a Multivariate Process  
Author Yuehjen E. Shao

 
Co-Author(s) Shih-Chieh Lin

 
Abstract When a disturbance or fault has been introduced into a process, the out-of-control signal would be triggered by the statistical process control (SPC) charts. However, for a multivariate process, the quality variables are responsible for this signal is difficult to be determined. Since accurately identify the faulty quality variables can significantly improve the process, the recognition of faulty quality variables has become a promising research topic. In this study, we propose three machine learning (ML) approaches to recognize the source of the fault for a multivariate process. These ML approaches include artificial neural network (ANN), extreme learning machine (ELM) and support vector machine (SVM). In this study, five quality variables is considered and the variance fault is presented in a multivariate normal process. A series of comput er experiments are conducted to evaluate the effectiveness of the proposed ML approaches.

 
Keywords Multivariate process, Variance fault, Machine learning, SPC
   
    Article #:  22334
 
Proceedings of the 22nd ISSAT International Conference on Reliability and Quality in Design
August 4-6, 2016 - Los Angeles, California, U.S.A.