A0678
Title: Sequential detection of drift changes in periodic mean-reverting processes
Authors: Yunhong Lyu - Trent University (Canada) [presenting]
Bouchra Nasri - University of Montreal (Canada)
Bruno N Remillard - HEC Montreal (Canada)
Abstract: A generalized Ornstein-Uhlenbeck process is introduced, tailored for modeling positive financial data exhibiting cyclic mean-reverting behavior. Unlike the classical O-U process, the formulation incorporates a periodic mean-reverting term, allowing the drift to fluctuate regularly over time. The central focus is on the sequential detection of change-points in the drift parameters of this process rather than traditional targets such as changes in mean, variance, or covariance. A set of detectors is developed to identify the timing of these changes and establish their asymptotic properties under both the null and alternative hypotheses. Results contribute both theoretical insights and practical tools for modeling dynamic financial systems with periodic structures.
