Investigating Continuous Optimization for fast Change Point Detection: Results of a short Research Internship at Macquarie University, Sydney.

09.01.2023, 10:15  –  Golm, Hs. 9, room 2.22 (and Zoom)
Forschungsseminar Statistik

Hans Reimann

Abstract:

Change point detection may be considered an essential step in data exploration and diagnostics. Offline approaches reformulating the problem as a subset selection problem in linear regression have been utillized before with some success, however, with new advances in continuous optimization for subset selection (COMBSS) new opportunities for fast change point detection in long sequences of data may arise. Next to first results and further steps, emerging issues and limitations will be discussed.

Zoom link on request