Proposal Description

Structural disorders in various fields, from materials science across economic universe to biological systems, present complex challenges in data analysis due to their inherent variability, various uncertainties, and changing patterns. Identification of structural breaks in data helps us to tie in specific health, legal, economic, or natural changes to the time epochs in which they occurred and the types of environments and industries that are affected. Testing and estimation of unknown changes hence reveals data segmentation, regime switching, and disorder recognition. Resampling methods, such as subsampling, bootstrapping and permutation techniques, offer computationally feasible as well as theoretically efficient strategies for assessing structural irregularities.

This session aims to explore the above mentioned resampling techniques in characterizing structural disorders; to examine their effectiveness in change-point analysis. Furthermore, the session is devoted to the implementation of resampling techniques in different domains - like irregularly spaced time series, multivariate counting processes, conformal prediction, deep learning, or online detection frameworks - affected by structural disorders, emphasizing their adaptability and usefulness across diverse disciplines; thereby fostering advancements in understanding and managing complex systems afflicted by such disorders.