Phase I outlier detection in profiles with binary data based on penalized likelihood. (11th September 2018)

• Record Type: Journal Article
• Authors: Li, Zhen; Shang, Yanfen; He, Zhen
• Is Part Of: Quality and reliability engineering international. Volume 35:Number 1(2019)
• Page Start: 1
• Page End: 13
• Publication Date: 2018-09-11
• Subjects: binary profile -- group LASSO -- masking effect -- outlier detection -- penalized likelihood; Reliability (Engineering) -- Periodicals; Quality control -- Periodicals; High technology -- Periodicals; 620.00452
• DOI: 10.1002/qre.2376
• Languages: English
• ISSNs: 0748-8017

Abstract

Abstract: Profile monitoring has been proven to be important among statistical process control problems. In some specific applications, the response variable of a profile can be categorical data, and numerous methods have been proposed for monitoring this type of profiles. Nevertheless, outlier detection for profiles with categorical data still attracted insufficient attention in the literature. To this end, this paper focuses on binary profiles and develops two schemes for outlier detection, from the viewpoint of penalized likelihood, based on the group LASSO method and directional information, respectively, in which the profiles of interest are treated as an integrated high‐dimensional vector. Simulation study shows that the proposed group‐type scheme usually performs better than the existing T I 2 chart in terms of detecting outliers correctly and alleviating the masking effect. Finally, a real example is used for illustrating the implementation of the proposed scheme.