Change-Point Models in Industrial Applications

This article discusses problems involving data streams subject to abrupt changes in time, such as shifts or drifts. Such data is typically described in terms of regimes and parameters that can be estimated, controlled or monitored. Two major problems in industrial setting are (a) detection of unfavorable changes and (b) estimation of the current level of parameters (filtering). The article discusses methods of addressing these problems within the framework of the change-point theory and gives examples illustrating their relevance in applications.