Detecting changes in data streams is an important task in several business and industrial applications, e.g. fraud detection in financial transactions or network intrusion in computer networks. Several approaches exist to detect changes, but an overview of state-of-the-algorithms is not easy to get because this is being studied in different scientic communities and thus various names are used (changepoint detection, anomaly detection, concept drift etc.). We will provide an overview of state-of-the-art approaches from both the statistical and data mining community so that business/industrial professionals know where to look for solutions for the challenges they face in detecting changes.

Alessandro Di Bucchianico is an Assistant Professor in the Department of Mathematics and Computer Science at Eindhoven University of Technology (TU/e). He is specialized in industrial statistics, statistical process control, reliability analyses, statistical computing, statistical software R and rare event simulation. You can find out more about him on his website.