Online Anomaly Detection of Profiles with Varying Coefficients via Functional Mixed Effects Modelling

Published: 18 November 2020| Version 1 | DOI: 10.17632/6p8npdvfn9.1
Contributors:
Dongdong Xiang,
Wendong Li

Description

In Xiang et al., we compare the efficiency of the proposed FMM control chart and three alternative algorithms for profile montioring by analyzing the real dataset from the tobacco manufacturing process. The dataset is provided by Shanghai tobacco group co., LTD. The entire dataset consists of 260 conforming profiles and 15 nonconforming profiles. In each profile, the response variable, the moisture content of tobacco leaf silk, and two covariates, the airflow temperature and the moisture content at the inlet, are recorded at 150 observed points. Based on the results, it can be concluded that the proposed FMM algorithm is the most effective for the online anomaly detection of profiles with varying coefficients.

Files