One-class classification-based control charts for multivariate process monitoring

摘要: One-class classification problems have attracted a great deal of attention from various disciplines. In the present study, attempts are made to extend scope application one-class technique Statistical Process Control (SPC) problems. New multivariate control charts that apply effectiveness improvement Phase I and II analysis in SPC proposed. These use monitoring statistic represent degree being an outlier as obtained through classification. The limits proposed established based on empirical level significance percentile, estimated by bootstrap method. A simulation study is conducted illustrate limitations current demonstrate charts.