Precision and Recall for Time Series

Abstract

Classical anomaly detection is principally concerned with point-based anomalies, those anomalies that occur at a single point in time. Yet, many real-world anomalies are range-based, meaning they occur over a period of time. Motivated by this observation, we present a new mathematical model to evaluate the accuracy of time series classification algorithms. Our model expands the well-known Precision and Recall metrics to measure ranges, while simultaneously enabling customization support for domain-specific preferences.

Cite

Text

Tatbul et al. "Precision and Recall for Time Series." Neural Information Processing Systems, 2018.

Markdown

[Tatbul et al. "Precision and Recall for Time Series." Neural Information Processing Systems, 2018.](https://mlanthology.org/neurips/2018/tatbul2018neurips-precision/)

BibTeX

@inproceedings{tatbul2018neurips-precision,
  title     = {{Precision and Recall for Time Series}},
  author    = {Tatbul, Nesime and Lee, Tae Jun and Zdonik, Stan and Alam, Mejbah and Gottschlich, Justin},
  booktitle = {Neural Information Processing Systems},
  year      = {2018},
  pages     = {1920-1930},
  url       = {https://mlanthology.org/neurips/2018/tatbul2018neurips-precision/}
}