Precision And Recall For Time Series

Authors:
Nesime Tatbul Intel Labs and MIT
Tae Jun Lee Microsoft
Stan Zdonik Brown University
Mejbah Alam Intel Labs
Justin Gottschlich Intel Labs

Introduction:

Classical anomaly detection is principally concerned with point-based anomalies, those anomalies that occur at a single point in time.Motivated by this observation, the authors present a new mathematical model to evaluate the accuracy of time series classification algorithms.

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.

You may want to know: