Online change-point detection for a transient change

TL;DR

This paper addresses the challenge of detecting transient changes in data distributions in real-time, reviewing existing methods, providing new approximations for false alarm rates, and comparing their effectiveness in normal data scenarios.

Contribution

It introduces new approximations for average run length to false alarm and compares the power of various procedures for transient change detection.

Findings

New approximations for false alarm rates

Comparison of detection procedures' power

Enhanced understanding of transient change detection performance

Abstract

We consider a popular online change-point problem of detecting a transient change in distributions of i.i.d. random variables. For this change-point problem, several change-point procedures are formulated and some advanced results for a particular procedure are surveyed. Some new approximations for the average run length to false alarm are offered and the power of these procedures for detecting a transient change in mean of a sequence of normal random variables is compared.