Detecting the Process Changes for Nonlinear Profile Data using SVR

• Main Authors: Chun-HanLiao, 廖俊翰
• Other Authors: Jeh-Nan Pan
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/g4z389

碩士 === 國立成功大學 === 統計學系 === 104 === In today’s manufacturing industries, if the quality characteristic of a product or a process is assumed to be represented by a functional relationship between the response variable and one or more explanatory variables, then the data generated from such a relationship is called profile data. Generally speaking, the functional relationship of the profile data can’t be known in advance and the real data usually are not follow normal distribution. Thus, the functional relationship of profile data is described via a non-parametric regression model, and a revised non-parametric EWMA control chart is proposed in the Phase II monitoring. In this research, we first fit the profile data through a support vector regression (SVR) model, then the fitted values are used to calculate the metrics. It is expected that the revised non-parametric EWMA control chart coupled with the metrics can be used for monitoring the profile data in the phase II study. Moreover, a simulations study is conducted to evaluate the detecting performance of the new control chart under various process shifts using average run length (ARL). Finally, a realistic nonlinear profile example is used to demonstrate the usefulness of our proposed non-parametric EWMA control chart and its monitoring schemes.