get_cps: Identify the Beginning of the Alarm Clusters

In VARcpDetectOnline: Sequential Change Point Detection for High-Dimensional VAR Models

Identify the Beginning of the Alarm Clusters

Description

This function clusters alarms into groups and identifies the starting points of the alarm clusters. If the next alarm occurs within a specified window size (w) from the current alarm, it will be considered part of the current cluster. Otherwise, a new cluster will be formed.

Usage

get_cps(alarms, w)

Arguments

alarms	A numeric vector. The alarms raised during the monitoring process.
w	An integer. The window size used to group alarms into clusters.

Value

A numeric vector containing the starting points of the alarm clusters. If the next alarm is within w observations of the current alarm, the next alarm will be considered part of the current alarm cluster. Otherwise, a new cluster is formed and the next alarm is considered the beginning of a new alarm cluster.

Examples

# Example usage:
alarms <- c(10, 15, 30, 35, 60)
change_points <- get_cps(alarms, w = 10)

VARcpDetectOnline documentation built on April 12, 2025, 1:44 a.m.