nab_artificialWithAnomaly: Numenta Anomaly Benchmark (NAB) – artificialWithAnomaly

In harbinger: A Unified Time Series Event Detection Framework

Numenta Anomaly Benchmark (NAB) – artificialWithAnomaly

Description

Synthetic time series with injected anomalies from the Numenta Anomaly Benchmark (NAB). Designed for evaluating anomaly detection algorithms under controlled conditions. Labels available: Yes. See cefet-rj-dal/united for detailed guidance on using this package and the other datasets available in it. Labels available? Yes

Usage

data(nab_artificialWithAnomaly)

Format

A list of time series.

Details

This package ships a mini version of the dataset. Use loadfulldata() to download and load the full dataset from the URL stored in attr(url).

Source

Numenta Anomaly Benchmark (NAB) Dataset

References

Lavin, A., & Ahmad, S. (2015). Evaluating real-time anomaly detection algorithms – the Numenta Anomaly Benchmark. 2015 IEEE 14th International Conference on Machine Learning and Applications (ICMLA).

Examples

data(nab_artificialWithAnomaly)
s <- nab_artificialWithAnomaly[[1]]
plot(ts(s$value), main = names(nab_artificialWithAnomaly)[1])

harbinger documentation built on Feb. 11, 2026, 9:06 a.m.