Restarted Bayesian Online Change-point Detector achieves Optimal Detection Delay

Part of Proceedings of the International Conference on Machine Learning 1 pre-proceedings (ICML 2020)

Bibtex »Metadata »Paper »Supplemental »

Authors

REDA ALAMI, Odalric-Ambrym Maillard, Raphaël Féraud

Abstract

In this paper, we consider the problem of sequential change-point detection where both the change-points and the distributions before and after the change are assumed to be unknown. For this key problem in statistical and sequential learning theory, we derive a variant of the Bayesian Online Change Point Detector proposed by \cite{adams2007bayesian} which is easier to analyze than the original version while keeping its powerful message-passing algorithm. We provide a non-asymptotic analysis of the false-alarm rate and the detection delay that matches the existing lower-bound. We further provide the first explicit high-probability control of the detection delay for such approach. Experiments on synthetic and real-world data show that this proposal compares favorably with the state-of-art change-point detection strategy, namely the Improved Generalized Likelihood Ratio (Improved GLR) while outperforming the original Bayesian Online Change Point Detection strategy.