Low-Count Time Series Anomaly Detection

• Published in: 2023 IEEE 33rd International Workshop on Machine Learning for Signal Processing (MLSP) pp. 1 - 6
• Main Authors: Renz, Philipp, Cutajar, Kurt, Twomey, Niall, Cheung, Gavin K. C., Xie, Hanting
• Format: Conference Proceeding
• Language: English
• Published: IEEE 17-09-2023
• Subjects: Analytical models; Anomaly detection; Benchmark testing; Intermittent Time Series; Low-Count Time Series; Monitoring; Smoothing methods; Surface treatment; Time series analysis; Time Series Anomaly Detection
• ISSN: 2161-0371
• DOI: 10.1109/MLSP55844.2023.10285979

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.