國立政治大學統計學系
學 術 演 講
主講人：Prof. Arthur B. Yeh (葉百堯教授)學 術 演 講
Department of Business Analytics, Economics and Information Systems
Bowling Green State University
題 目：Multivariate Control Charts: An Overview of Past and Recent Developments
時 間：民國115年6月9日 (星期二) 上午9：10~12：00
地 點：國立政治大學商學院261145教室
摘 要：
In this lecture, I will give an overview of some of the developments of multivariate control charts in the last three decades. The aim is to discuss these developments in a unified framework, specifically with respect to Shewhart charts vs. CUSUM and EWMA charts, Phase I control vs. Phase II monitoring, monitoring the mean vector vs. monitoring the covariance matrix, etc. I will then discuss some recent developments in multivariate control charts, especially with respect to multiple testing procedure based multivariate control chart for monitoring process mean. We will discuss in more detail how multiple testing procedures can be adapted to construct multivariate control charts. The control charts thus developed can not only monitor correlated variables but also provide instantaneous diagnostics of out-of-control parameters when an out-of-control signal is detected, without much further efforts. In essence, the multiple testing procedures-based multivariate control charts simultaneously tackle and provide solutions to two major challenges, monitoring/detection and diagnostics, in multivariate control chart developments and applications. Furthermore, these types of charts offer potential solutions to problems that have received little or almost no attention in existing literature, e.g., multivariate control charts for monitoring correlated variables of different types.
Keywords: Monitoring and Detection; Multiple testing procedures; Multivariate control charts.
