harbinger (version 1.2.747)

hanr_arima: Anomaly detector using ARIMA

Description

Fits an ARIMA model to the series and flags observations with large model residuals as anomalies. Wraps ARIMA from the forecast package.

Usage

hanr_arima()

Value

hanr_arima object.

Details

The detector estimates ARIMA(p,d,q) and computes standardized residuals. Residual magnitudes are summarized via a distance function and thresholded with outlier heuristics from harutils().

References

• Box GEP, Jenkins GM, Reinsel GC, Ljung GM (2015). Time Series Analysis: Forecasting and Control. Wiley.

Examples

library(daltoolbox)

# Load anomaly example data
data(examples_anomalies)

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

# Configure ARIMA anomaly detector
model <- hanr_arima()

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

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

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