Precise Change Point Detection using Spectral Drift Detection

TL;DR

This paper introduces a spectral drift detection method that leverages kernel embedding spectral properties to accurately identify change points in data distributions over time, improving robustness over existing approaches.

Contribution

It proposes a novel spectral-based unsupervised change point detection algorithm utilizing kernel embeddings, with theoretical analysis and experimental validation.

Findings

Effective detection of multiple drift events

Reduced false positives in small windows

Strong theoretical properties and practical performance

Abstract

The notion of concept drift refers to the phenomenon that the data generating distribution changes over time; as a consequence machine learning models may become inaccurate and need adjustment. In this paper we consider the problem of detecting those change points in unsupervised learning. Many unsupervised approaches rely on the discrepancy between the sample distributions of two time windows. This procedure is noisy for small windows, hence prone to induce false positives and not able to deal with more than one drift event in a window. In this paper we rely on structural properties of drift induced signals, which use spectral properties of kernel embedding of distributions. Based thereon we derive a new unsupervised drift detection algorithm, investigate its mathematical properties, and demonstrate its usefulness in several experiments.