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.
@article{arxiv.1803.03639,
title = {Precision and Recall for Time Series},
author = {Nesime Tatbul and Tae Jun Lee and Stan Zdonik and Mejbah Alam and Justin Gottschlich},
journal= {arXiv preprint arXiv:1803.03639},
year = {2019}
}
Comments
11 pages, 32nd Conference on Neural Information Processing Systems (NeurIPS 2018), Montreal, Canada