| Summary: | 博士 === 國立臺灣科技大學 === 工業管理系 === 100 === In manufacturing process, a sequence of measurements of quality characteristic is increasingly taken across some continuum, producing a curve that represents the quality of the item. This curve provides the so-called functional data. Traditionally, functional data is treated as a special type of profile data. Regardless of a linear or nonlinear profile, the common approaches of the control chart are based on the multivariate control chart by monitoring the estimated parameter of the predefined linear or nonlinear model. Usually, the model is difficult to know practically, and it is also difficult to identify the abnormal pattern from the outlying parameter. By using the techniques of functional data analysis, we proposed the functional data control chart which can provide a better solution to these problems. In the Monte Carlo simulations, we show that the functional data control chart is sensitive when the underlying process status is changed. By applying to real example data, the new method exhibits a good performance.
|
|---|
