Low-count Time Series Anomaly Detection

TL;DR

This paper introduces a new benchmark dataset and analysis for low-count time series anomaly detection, highlighting challenges and proposing smoothing techniques to improve detection performance in sparse, real-world data.

Contribution

The paper presents a novel generative benchmark for low-count time series with anomalies and provides insights and methods to enhance anomaly detection in such challenging settings.

Findings

Widely-used algorithms struggle with distribution overlap in low-count data.

Anomaly score smoothing improves detection performance.

The approach is validated on real-world retail sales data.

Abstract

Low-count time series describe sparse or intermittent events, which are prevalent in large-scale online platforms that capture and monitor diverse data types. Several distinct challenges surface when modelling low-count time series, particularly low signal-to-noise ratios (when anomaly signatures are provably undetectable), and non-uniform performance (when average metrics are not representative of local behaviour). The time series anomaly detection community currently lacks explicit tooling and processes to model and reliably detect anomalies in these settings. We address this gap by introducing a novel generative procedure for creating benchmark datasets comprising of low-count time series with anomalous segments. Via a mixture of theoretical and empirical analysis, our work explains how widely-used algorithms struggle with the distribution overlap between normal and anomalous segments. In order to mitigate this shortcoming, we then leverage our findings to demonstrate how anomaly score smoothing consistently improves performance. The practical utility of our analysis and recommendation is validated on a real-world dataset containing sales data for retail stores.