Changepoint Detection

What's Our Research Expertise?

  • Changepoints are sudden - often unexpected - shifts in the behaviour of a physical, biological, industrial or financial process. Why is it important to identify these shifts?
  • In climate, they can help improve understanding of climate change and the impacts of emissions policies on climate.
  • They can be a warning signal for interventions in an industrial setting, or the need for treatment in a medical one.
  • In communications, they can flag unexpected behaviour which may be linked to hacking.
  • Efficient change-detection can either be
    • a stand-alone tool to recognise that something has changed in an underlying system, or
    • a component of a larger system for accurate and efficient prediction.
  • Lancaster University is a world-leading centre of excellence in innovative changepoint detection methodology.
  • Our work has been used in retail, finance, digital health, space exploration, environmental science, cybersecurity and engineering.
  • We collaborate with industry including BT and Tesco to deploy changepoint methods to monitor core business processes.
  • NASA use our methods to monitor for changes in the air quality in space suits!

What Are We Teaching?

Change-point detection is rarely taught at undergraduate level but our new-for-24/25 third-year module MATH337 will do just this! In this module, you will experience applications, theory, algorithms and practical software implementations. The module builds on fundamental probability (MATH230) and statistics (MATH235), and links well to our third-year modules on Time-Series Analysis (MATH334) and the Mathematics of AI (MATH336).

In this module you will develop skills for modern data challenges, and the module is an ideal learning experience and strong foundation for anybody wishing to develop a career in AI, data science or statistics.

Mathematics Making a Difference