Research Article
A two-stage monitoring scheme for high-dimensional Poisson data
@INPROCEEDINGS{10.4108/eai.24-2-2023.2330695, author={Huantong Na and Xuemin Zi}, title={A two-stage monitoring scheme for high-dimensional Poisson data}, proceedings={Proceedings of the 2nd International Conference on Engineering Management and Information Science, EMIS 2023, February 24-26, 2023, Chengdu, China}, publisher={EAI}, proceedings_a={EMIS}, year={2023}, month={6}, keywords={high-dimensional poisson data cusum control chart arl fdr}, doi={10.4108/eai.24-2-2023.2330695} }- Huantong Na
Xuemin Zi
Year: 2023
A two-stage monitoring scheme for high-dimensional Poisson data
EMIS
EAI
DOI: 10.4108/eai.24-2-2023.2330695
Abstract
In the field of industrial quality control, when monitoring continuous data is mainly considered, it is also necessary to monitor discrete data, among which high-dimensional Poisson distribution data is very common in modern manufacturing. At present, a large number of the literature proposed to use of statistical process technology to monitor high-dimensional data, but the scheme for monitoring high-dimensional Poisson data in the existing literature is rare, and the few proposed methods for high-dimensional Poisson data to monitor the false discovery rate (FDR) of data were given, which is popular statistic for constructing the control chart. In order to effectively monitor the high-dimensional Poisson data, this paper extended the methodology of (Li, 2018) proposes a two-stage monitoring algorithm, which uses the CUSUM control chart to control the in-control (IC) average run length (ARL) in the first stage, and uses pointwise FDR to control the Type-I error rate in the second stage. In this paper, numerical simulation is used to realize the monitoring process and the performance demonstrates the efficiency and robustness of the proposed procedure.