QPRC 2016

Nonparametric Dynamic Screening System for Monitoring Correlated Longitudinal Data

Jun Li
Department of Statistics
University of California
Riverside, CA


In many applications, including disease early detection and prevention, and performance evaluation of airplanes and other durable products, we need to sequentially monitor the longitudinal pattern of certain performance variables of a subject. A signal should be given as soon as possible once the pattern becomes abnormal. Recently, a new statistical method called dynamic screening system (DySS) has been proposed to solve this problem. It is a combination of longitudinal data analysis and statistical process control. However, the current DySS method can only handle cases when observations are normally distributed and within-subject observations are independent or follow a specific time series model (e.g., AR(1) model). In this talk, we propose a new nonparametric DySS method which can handle cases when the observation distribution and the correlation among within-subject observations are arbitrary. Therefore, it broadens the application of the DySS method greatly. Numerical studies show that the new method works well in practice.