hcp_pelt: Pruned Exact Linear Time (PELT)

In harbinger: A Unified Time Series Event Detection Framework

Pruned Exact Linear Time (PELT)

Description

Multiple change-point detection using the PELT algorithm for mean/variance with a linear-time cost under suitable penalty choices. This function wraps the PELT implementation in the changepoint package.

Usage

hcp_pelt()

Details

PELT performs optimal partitioning while pruning candidate change-point locations to achieve near-linear computational cost.

Value

hcp_pelt object.

References

• Killick R, Fearnhead P, Eckley IA (2012). Optimal detection of changepoints with a linear computational cost. JASA, 107(500):1590–1598.

Examples

library(daltoolbox)

# Load change-point example data
data(examples_changepoints)

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

# Configure the PELT detector
model <- hcp_pelt()

# Fit the detector (no-op for PELT)
model <- fit(model, dataset$serie)

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

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

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