A Study on the Multivariate Combined Control Chart

• 발행기관: 한국자료분석학회
• 저자명: 홍승만(Seung-Man Hong)
• 간행물 정보: 『Journal of The Korean Data Analysis Society (JKDAS)』Vol.21 No.5, 2193~2202쪽, 전체 10쪽
• 주제분류: 자연과학 > 통계학
• 파일형태: PDF
• 발행일자: 2019.10.31

영문 초록

In this research, we propose multivariate combined control charts for the mean vector and covariance matrix to control simultaneously in the process under the normality assumption. For doing this, first of all, we review the multivariate simultaneous tests proposed by Park (2018) in some detailed fashion, which makes use of the likelihood ratio principle. Also we consider the simultaneous tests with various combining functions to combine individual tests. Then we propose to construct the multivariate combined charts by investigating the relation between the hypothesis test and control chart to utilize the multivariate simultaneous tests. We note that the proposed control charts will be consisted with a series of hypothesis tests instead of plotting the values of sample mean vector and covariance matrix in the control chart sheets. Then we investigate the efficiency among the proposed control charts by obtaining empirical average run length which represents the average runs until detecting out of control state when the process is out of control. Finally, we observe some interesting features related with the multivariate combined control charts.

목차

1. Introduction

2. Review of the simultaneous tests

3. Multivariate combined charts

4. A simulation study and concluding remarks

References

키워드

average run length combining function combined control chart likelihood ratio principle multivariate normal distribution