What curves can reveal about complex data

Some data are better understood as curves than as individual numbers. Blood pressure or heart-rate measurements collected repeatedly over a day, for example, form a trajectory whose shape can be just as informative as any single reading.
For statisticians, such functional data raise difficult questions. How can entire curves be compared across groups? What happens when observations are missing, sample sizes are small, or standard assumptions about the data do not hold?
These are the problems addressed by OUTCLASS – Powerful Inference for Functional Data in Complex Factorial Designs. The joint WEAVE NCN-DFG project develops robust statistical methods for functional data in settings where several factors may influence the observations. The methods are intended for medical application context and also applicable for other research fields including  psychology, educational sciences, or biology, where complex measurement trajectories occur frequently.

From theory to possible applications

From September 28 to October 1, 2026, Markus Pauly and Marléne Baumeister visited project partners Natalia Stefańska and Prof. Łukasz Smaga at Adam Mickiewicz University in Poznań. The meeting brought the German and Polish teams together for several days of concentrated work.
One part of the visit was devoted to the statistical theory behind OUTCLASS. The extended in-person exchange gave the researchers time to discuss mathematical questions in greater depth and prove theoretical aspects together. For Marléne Baumeister, the meeting also provided an opportunity to explore these questions directly with both PIs (Łukasz Smaga and Markus Pauly).
The project team also exchanged ideas with researchers who work with functional data in practice, including one colleague who joined virtually from Iceland. These discussions focused on possible collaborations and on data settings in which the methods developed within OUTCLASS could eventually be tested or applied.

A Polish-German collaboration

OUTCLASS is led by the two Principal Investigators Prof. Łukasz Smaga at Adam Mickiewicz University and Prof. Markus Pauly at TU Dortmund University. Markus Pauly heads the Chair of Mathematical Statistics and Applications in Industry at TU Dortmund University and is a Principal Investigator at RC Trust. Marléne Baumeister is a researcher in his team and contributes to OUTCLASS. She has also been involved in interdisciplinary research at RC Trust, including work on trust in and adoption of statistical methods.
The visit in Poznań also looked ahead. The partners began planning a project-related workshop that is expected to take place in Dortmund in 2027 and continue the exchange between the teams and researchers working with functional data.

About the project

OUTCLASS – Powerful Inference for Functional Data in Complex Factorial Designs is a joint WEAVE project supported on the German side by the German Research Foundation (DFG) and on the Polish side by the National Science Centre (NCN). The collaboration connects statistical research at TU Dortmund University and Adam Mickiewicz University in Poznań.