Greedy online change point detection

TL;DR

This paper introduces GOCPD, a computationally efficient online change point detection method that maximizes data likelihood under model independence, reducing false positives and validated on synthetic and real data.

Contribution

The paper presents a novel greedy online change point detection algorithm that is faster and more robust against outliers than existing methods.

Findings

Unimodal objective for single change points enables efficient ternary search.

GOCPD outperforms traditional methods on synthetic data.

Validated effectiveness on real-world univariate and multivariate datasets.

Abstract

Standard online change point detection (CPD) methods tend to have large false discovery rates as their detections are sensitive to outliers. To overcome this drawback, we propose Greedy Online Change Point Detection (GOCPD), a computationally appealing method which finds change points by maximizing the probability of the data coming from the (temporal) concatenation of two independent models. We show that, for time series with a single change point, this objective is unimodal and thus CPD can be accelerated via ternary search with logarithmic complexity. We demonstrate the effectiveness of GOCPD on synthetic data and validate our findings on real-world univariate and multivariate settings.