Adaptive Bernstein Change Detector for High-Dimensional Data Streams

TL;DR

The paper introduces ABCD, an adaptive change detector for high-dimensional data streams that identifies change points, subspaces, and severity levels by monitoring model accuracy with a Bernstein-based score.

Contribution

ABCD is a novel change detection method that learns an encoder-decoder model and adaptively monitors its accuracy to detect and characterize changes in high-dimensional data streams.

Findings

ABCD outperforms competitors by up to 20% in F1-score.

It accurately estimates change subspaces.

It provides a severity measure correlated with ground truth.

Abstract

Change detection is of fundamental importance when analyzing data streams. Detecting changes both quickly and accurately enables monitoring and prediction systems to react, e.g., by issuing an alarm or by updating a learning algorithm. However, detecting changes is challenging when observations are high-dimensional. In high-dimensional data, change detectors should not only be able to identify when changes happen, but also in which subspace they occur. Ideally, one should also quantify how severe they are. Our approach, ABCD, has these properties. ABCD learns an encoder-decoder model and monitors its accuracy over a window of adaptive size. ABCD derives a change score based on Bernstein's inequality to detect deviations in terms of accuracy, which indicate changes. Our experiments demonstrate that ABCD outperforms its best competitor by up to 20% in F1-score on average. It can also accurately estimate changes' subspace, together with a severity measure that correlates with the ground truth.