Optimal Sub-sampling to Boost Power of Kernel Sequential Change-point Detection

TL;DR

This paper introduces an optimal sub-sampling method to enhance the detection power of kernel-based sequential change-point detection methods, effectively addressing power loss due to large historical data.

Contribution

The paper proposes a novel optimal sub-sampling scheme for kernel change-point detection, improving detection power in large data settings.

Findings

Enhanced detection performance demonstrated through numerical experiments

Optimal sub-sampling reduces power loss in large datasets

Applicable to Scan B and Kernel CUSUM procedures

Abstract

We present a novel scheme to boost detection power for kernel maximum mean discrepancy based sequential change-point detection procedures. Our proposed scheme features an optimal sub-sampling of the history data before the detection procedure, in order to tackle the power loss incurred by the random sub-sample from the enormous history data. We apply our proposed scheme to both Scan $B$ and Kernel Cumulative Sum (CUSUM) procedures, and improved performance is observed from extensive numerical experiments.