TY - JOUR
T1 - Efficient monitoring of autocorrelated Poisson counts
AU - Li, Jian
AU - Zhou, Qiang
AU - Ding, Dong
N1 - Funding Information:
Dr. Li’s research was supported by the National Natural Science Foundation of China under grants 71772147, 71402133, and 71572138; the Fundamental Research Funds for the Central Universities; the Youth Innovation Team of Shaanxi Universities “Big data and Business Intelligent Innovation Team.” Dr. Ding’s research was supported by the National Natural Science Foundation of China under grant 71602155. The authors would like to thank the Department Editor and two anonymous referees for their many helpful comments that have resulted in significant improvements in this article.
Publisher Copyright:
© 2019, © 2019 “IISE”.
PY - 2020/7/2
Y1 - 2020/7/2
N2 - Statistical surveillance for autocorrelated Poisson counts has drawn considerable attention recently. These works are usually based on a first-order integer-valued autoregressive model and focus on monitoring separately either the marginal mean or the autocorrelation coefficient. Inspired by multivariate statistical process control, this article transforms autocorrelated Poisson counts into a bivariate representation and proposes an efficient control chart. By borrowing the power of the likelihood ratio test, albeit surprisingly, this chart demonstrates almost uniformly stronger power than the existing alternatives in simultaneously detecting shifts in both the marginal mean and the autocorrelation coefficient. In addition, the robustness of the proposed chart against overdispersion encountered often in counts is also verified. It is shown that this chart also has superiority in monitoring autocorrelated overdispersed counts.
AB - Statistical surveillance for autocorrelated Poisson counts has drawn considerable attention recently. These works are usually based on a first-order integer-valued autoregressive model and focus on monitoring separately either the marginal mean or the autocorrelation coefficient. Inspired by multivariate statistical process control, this article transforms autocorrelated Poisson counts into a bivariate representation and proposes an efficient control chart. By borrowing the power of the likelihood ratio test, albeit surprisingly, this chart demonstrates almost uniformly stronger power than the existing alternatives in simultaneously detecting shifts in both the marginal mean and the autocorrelation coefficient. In addition, the robustness of the proposed chart against overdispersion encountered often in counts is also verified. It is shown that this chart also has superiority in monitoring autocorrelated overdispersed counts.
KW - Autocorrelation coefficient
KW - bivariate Poisson distribution
KW - marginal distribution
KW - overdispersion
KW - statistical process control
UR - http://www.scopus.com/inward/record.url?scp=85071316017&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85071316017&partnerID=8YFLogxK
U2 - 10.1080/24725854.2019.1649506
DO - 10.1080/24725854.2019.1649506
M3 - Article
AN - SCOPUS:85071316017
SN - 2472-5854
VL - 52
SP - 769
EP - 779
JO - IIE Transactions (Institute of Industrial Engineers)
JF - IIE Transactions (Institute of Industrial Engineers)
IS - 7
ER -