hanr_fft: Anomaly detector using FFT

In harbinger: A Unified Time Series Event Detection Framework

Anomaly detector using FFT

Description

High-pass filtering via FFT to isolate high-frequency components; anomalies are flagged where the filtered magnitude departs strongly from baseline.

Usage

hanr_fft()

Details

The spectrum is computed by FFT, a cutoff is selected from the power spectrum, low frequencies are nulled, and the inverse FFT reconstructs a high-pass signal. Magnitudes are summarized and thresholded using harutils().

Value

hanr_fft object

References

• Sobrinho, E. P., Souza, J., Lima, J., Giusti, L., Bezerra, E., Coutinho, R., Baroni, L., Pacitti, E., Porto, F., Belloze, K., Ogasawara, E. Fine-Tuning Detection Criteria for Enhancing Anomaly Detection in Time Series. In: Simpósio Brasileiro de Banco de Dados (SBBD). SBC, 29 Sep. 2025. doi:10.5753/sbbd.2025.247063

Examples

library(daltoolbox)

# Load anomaly example data
data(examples_anomalies)

# Use a simple example
dataset <- examples_anomalies$simple
head(dataset)

# Configure FFT-based anomaly detector
model <- hanr_fft()

# Fit the model
model <- fit(model, dataset$serie)

# Run detection
detection <- detect(model, dataset$serie)

# Show detected anomalies
print(detection[(detection$event),])

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